EP 158 - Rafael Rosengarten: Why 90% of Cancer Drugs Fail — and the Radical AI Fix You’ve Never Heard Of

 Most cancer drugs fail. Not because the science is wrong—because we’re solving the wrong problems.

The cost? Over $2 billion per failure. And for the patient waiting on a miracle—there’s no second chance.

Behind the headlines of “precision medicine,” there’s a deeper story nobody’s telling. Until now. 

🎯 Enter Rafael Rosengarten, the scientist-turned-founder who’s rewriting the rules of drug development.

In this gripping conversation, we unpack how RNA, AI, and deep empathy could finally close the loop between biology, data, and the patient in the room. 

🎧 Watch now to explore:
1️⃣ Why drug failure isn’t a tech problem—it’s a strategy problem
2️⃣ How “information companions” will guide every medicine to the right patient
3️⃣ What pharma keeps getting wrong about biomarkers—and how to fix it
4️⃣ The untold story behind turning down a chef’s job on a Mediterranean yacht
5️⃣ The real reason Genialis may one day go out of business (and why Rafael hopes it does) 

👤 About Rafael Rosengarten
Rafael is CEO and co-founder of Genialis, the RNA biomarker company reshaping precision oncology. From academic labs to Michelin-star kitchens, his journey is anything but linear—but his mission is clear: make medicine make sense. With partners across pharma, diagnostics, and AI, Genialis is creating a world where every patient gets the treatment they actually need. 

💬 Quotes That Might Shift Your Thinking:
(00:58:37) “Our AI biomarkers de-risk clinical trials, slashing costs and saving lives.”
(01:54:46) “Every drug will have an information companion guiding it to the right patient.”
(01:14:32) “AI should free humans for empathy and creativity, not replace them.”
(00:14:13) “We now have the tools to treat every cancer as unique to each patient.”
(01:31:38) “Stop selling technology; sell solutions to real problems.” 

🧭 Timestamps to Explore:
(00:03:52) Inside the World’s Largest Medical Center
(00:10:30) How AI Is Rewiring Precision Medicine
(00:13:51) Why Precision Medicine Is Scaling Now
(00:17:34) Startups Must Solve Market Needs First
(00:23:53) Culture First, Then Business: Startup Blueprint
(00:27:57) Cancer Isn’t One Disease—It’s Thousands
(00:31:44) Predicting Drug Success with AI Biology
(00:36:40) Reimagining Biomarkers Using RNA and AI
(00:47:45) Why Early Detection Still Saves Lives
(00:57:30) Why Most Cancer Drugs Still Fail
(00:58:37) AI Biomarkers That De-Risk Clinical Trials
(01:09:42) Rethinking Dosage with Predictive Algorithms
(01:13:33) Why Human Intuition Still Matters in Science
(01:36:41) The Crystal Ball Question Pharma Can’t Ignore
(01:50:19) The moment where time disappears—and the real future begins.
(01:53:39) Every Drug Will Have an AI Companion 

🔔 Let’s Rebuild Trust in Medicine—Together
If this episode sparked a new thought—or just gave you pause—don’t let it end here.

Subscribe. Share it with someone who still believes medicine can be both scientific and deeply human.

Every action brings in minds like Rafael’s… and gets us one step closer to treatments that truly heal. 

🎧 Watch now—and discover how medicine gets smarter, safer, and more personal. 

Transcript

00:00:00:00 - 00:00:07:12

Christian Soschner

Everyone talks about curing cancer, but nobody talks about the 90% of drugs that fail

00:00:07:12 - 00:00:12:05

Christian Soschner

drug development. It's the costliest problem in medicine.

00:00:12:05 - 00:00:24:24

Christian Soschner

Up to $5 billion lost Protract decades wasted lives cut short and worse. We often don't even know why it felt

00:00:24:24 - 00:00:26:14

Rafael Rosengarten

We think we can do better by a lot,

00:00:26:14 - 00:00:36:09

Christian Soschner

The real revolution. It's not a drug. It's the algorithm behind the drug. Keneally's doesn't make medicine.

00:00:36:11 - 00:00:47:17

Christian Soschner

They make medicine make sense by translating RNA into insight. They're making every treatment smarter, faster, and more human.

00:00:47:17 - 00:00:52:16

Rafael Rosengarten

every drug is going to have kind of an information companion that helps it on its journey

00:00:52:16 - 00:00:53:23

Christian Soschner

And the men behind it.

00:00:53:23 - 00:01:04:08

Christian Soschner

The scientist who once turned down a chef's chop on a mediterranean yacht because he believed a harder, lonelier road might just save more lives.

00:01:04:08 - 00:01:06:18

Rafael Rosengarten

I actually think about that decision every day of my life,

00:01:06:18 - 00:01:14:21

Christian Soschner

But before we dive in, just one ask if this episode shifts your thinking even just a little.

00:01:14:23 - 00:01:25:21

Christian Soschner

Follow the show. Leave a review or share it with someone you respect. That's how we bring in more grad classmates and take these conversations even deeper.

00:01:25:21 - 00:01:33:09

Christian Soschner

This isn't just an episode, it's an invitation to rewire how we build, test, and trust medicine.

00:01:33:09 - 00:01:44:06

Christian Soschner

But I feel good to see you. We were just talking about, Houston by Houston is so special. Why does. Why are 150 languages spoken in Houston? What's the reason?

00:01:44:08 - 00:01:48:18

Rafael Rosengarten

Oh, gosh. I'm probably going to give you an apocryphal answer, but the reason that I was told

00:01:48:18 - 00:02:05:18

Rafael Rosengarten

is that, this is what the major port of entry for immigrants coming into the United States, especially under, refugee status or asylum status or any other government mediated relocation program. And of course, you know, all of those sorts of things are in peril today.

00:02:05:18 - 00:02:31:00

Rafael Rosengarten

But historically speaking, this is where people just land first and often, you know, there's already a community you can stick around. I used to do volunteer work at one of the Houston refugee centers. This was in the kind of 2018, so 2011 to 2015. And there were a huge number of newly relocated Middle Eastern, South Asian refugees, you know, that were displaced by the ongoing wars.

00:02:31:01 - 00:02:38:09

Rafael Rosengarten

And they would come here and there were already, you know, communities, mosques, restaurants, you know, things like this and would often stick.

00:02:38:21 - 00:02:44:11

Christian Soschner

You is the first port of entry or the first point of entry for immigrants.

00:02:44:14 - 00:03:01:02

Rafael Rosengarten

This is again, this could be complete nonsense. It's just what I had heard when I asked somebody. So when when the US government would relocate, or was involved in the relocation of an immigrant, this is often the port they would come through. Don't take me at my word. Google it. I'm sure there's a real answer out there somewhere.

00:03:01:05 - 00:03:21:24

Rafael Rosengarten

I mean, it remains true that it's an incredibly diverse city, an incredibly, plural city in a very functional way. Like in your everyday life. You run into people from all over the world. I would argue there might be another reason the major industries here are oil and gas and medicine, and both oil and gas and medicine tend to be, sort of geographically rather diverse.

00:03:21:24 - 00:03:28:04

Rafael Rosengarten

Right? Oil and gas is a global industry, a global trade. And, you know, medicine really doesn't know boundaries.

00:03:28:09 - 00:03:33:23

Christian Soschner

Well, why is medicine so famous in Houston? What's, what's the main point that you want to tell the audience?

00:03:33:23 - 00:03:34:19

Rafael Rosengarten

Yeah, the Texas

00:03:34:19 - 00:03:54:11

Rafael Rosengarten

Medical Center in Houston is the world's largest contiguous medical center. I don't know how many institutions make it up today, but when I was a researcher there, again, 2011 to 2015, there were 69 or 70 partner institutions, shoulder to shoulder to shoulder. A good friend of mine came to visit for a weekend, and I took him to show him around the medical center.

00:03:54:11 - 00:04:14:10

Rafael Rosengarten

And he looks up, staring at these towering glass hospitals and research institutes. And this man, this is like Las Vegas for hospitals. And it really is. So the Texas Medical Center is this extraordinary place. But of course, it had to start from somewhere. And Baylor College of Medicine was one of the anchor institutions. That's where I did my postdoctoral training.

00:04:14:12 - 00:04:39:06

Rafael Rosengarten

But there are some quite famous places, like MD Anderson Cancer Center. You know, it's arguably the most famous cancer center in the world. Is there? It's quite famous for, cardiology. DeBakey, who is a famous heart surgeon who invented most of what we think of as modern heart surgery. He was there and in turn, world leaders, you know, famous presidents and prime ministers and autocrats would come here for their lifesaving therapies.

00:04:39:06 - 00:04:41:15

Rafael Rosengarten

And it really kind of put the place on the map.

00:04:41:20 - 00:04:53:06

Christian Soschner

Oh, that's great. I didn't know that the, the signal centers are in use in Texas. It's always good to learn something new by podcasting. And this is the reason why you relocated. Tend to Houston and started your company there?

00:04:53:06 - 00:05:16:19

Rafael Rosengarten

Know? Quite the contrary. I'm smart and I married well. And so my wife had while I was playing at being an academic and then an early startup guy. She had a real job as a real job. So she works for a big, multinational energy company. And so we moved to Houston, for her work. We had been previously in the Bay area, and before that I was, in the northeast, in the United in New Haven, Connecticut, in the US.

00:05:16:21 - 00:05:34:14

Rafael Rosengarten

So I followed her kind of to California and then to Houston following her work. But it turns out if you're a scientist, Houston is a perfectly good place to land if you need to do science. Because, again, we have the world's largest medical center. So finding a really top quality lab to continue my my postdoctoral work was was not so hard.

00:05:34:14 - 00:05:45:16

Christian Soschner

You have your blast. Basically, you lived in, San Francisco in the Bay area. And then I mean, Houston, when you look at entrepreneurship and investing in both areas, what makes it so special?

00:05:45:19 - 00:06:06:12

Rafael Rosengarten

Yeah. Well, the Bay area has an unparalleled ecosystem for for entrepreneurship, you know, in the life sciences. Boston, the general Boston areas. Okay. San Diego's okay, but the Bay area has this incredible ecosystem is the right word because it it needs at least two pieces. It needs both the the novel technology and entrepreneurial spirit and deal flow. Right.

00:06:06:12 - 00:06:34:05

Rafael Rosengarten

New companies. But it also needs investors. And it's a bit of a chicken and egg problem. If you're trying to build a new startup hub, how do you attract investors if there aren't so many startups? And how do you get people to take a risk and build a startup if there's no way to fund it? And so since I've been associated with Houston starting in 2011, and then when I began my entrepreneurial journey in 2015, I have witnessed the growth of an ecosystem from very much.

00:06:34:05 - 00:06:54:11

Rafael Rosengarten

What was a cold start problem? You needed companies and you needed investors, and you needed to figure out how to grow them both at once. The Houston is now a much more competitive entrepreneurial ecosystem than it was 14, 15 years ago. Like, you know, night and day different. But we still need more investors and we still need more startups.

00:06:54:11 - 00:07:16:22

Rafael Rosengarten

And so there's a bit of a pattern here where we can build really, really promising early stage companies. And again, we have all this access to the hospitals, to the medical center, top research, top innovation and IP. But we don't we lack the community of seasoned operators. So when companies grow and they go to raise maybe their first baby or second big institutional venture round, they very often move to the coasts.

00:07:16:24 - 00:07:35:08

Rafael Rosengarten

They'll go to Boston or they'll go to the West Coast seeking not only investment but more seasoned operators. It doesn't happen always. There's some great companies built in born here that are built here. But even genealogy is a bit like that. We started here in Houston. I went out to the Bay area, I then actually traveled overseas, and now I'm back in Houston.

00:07:35:10 - 00:07:41:13

Rafael Rosengarten

And all the while we've built our team where the talent was rather than focusing on one one geography.

00:07:41:17 - 00:07:50:19

Christian Soschner

That's, that's a good thing. So you ran basically. A virtual company. It's not only located in Houston, Texas, but also in other cities.

00:07:50:22 - 00:08:00:18

Rafael Rosengarten

So we have two offices with permanent seats. We have a commercial office in Boston, and then we have our larger office in Ljubljana, Slovenia. And Ljubljana is where my co-founder

00:08:00:18 - 00:08:04:19

Rafael Rosengarten

is from. And I don't know how many desks exactly we have, but it's like a

00:08:04:19 - 00:08:10:14

Rafael Rosengarten

25% off or something like this. We're 35 people globally. So it's the it's the bulk of the workforce.

00:08:10:16 - 00:08:34:17

Rafael Rosengarten

So it's it is virtual in a sense. I'm remote to the team, but we very much have a physical presence. Right. And we have a place for people to have the kind of synergies and the, the, you know, serendipity you get from sharing, sharing a workspace with others. But, you know, I myself don't get to take advantage of that very often, only when I travel over overseas or up to Boston to be with the rest of the team.

00:08:34:23 - 00:08:38:10

Christian Soschner

Now, I'm really tempted to ask one question.

00:08:38:12 - 00:08:39:06

Rafael Rosengarten

Yeah, sure. You have.

00:08:39:06 - 00:09:01:03

Christian Soschner

Experience. San Francisco. You're familiar with Texas, also Boston and, the other areas in the United States and and Europe and Europe with your co-founder in Slovenia, when you look at the European ecosystem, what makes the European ecosystem special for you, that you say it makes sense for your company to stay in Europe and not pull everything over to the United States?

00:09:01:05 - 00:09:24:18

Rafael Rosengarten

It's a great question. So we have a lot of Slovenian roots. First of all, the company does, the origin story is the genealogy originally founded by three Slovenian colleagues, a data scientist, another I.T guy who was really interested in sort of building a scalable company. And a third musketeer who I think obviously by swingman, he would do everything, anything that was needed.

00:09:24:20 - 00:09:48:11

Rafael Rosengarten

Really versatile guy. So that was Slovenia. I actually met these guys, guys through the research. Right. So I was a scientist at at Baylor College of Medicine here in Houston. My co-founder Misha stage to he's still our chief technology officer. He came over on what we could call kind of an industry postdoc. Right. He had finished his PhD and I, I was the biologist UCI guy, and he wanted to start this company.

00:09:48:11 - 00:10:11:19

Rafael Rosengarten

And his professor, his PhD advisor, had an appointment at Baylor and said, well, take your technology, go to Baylor and ask everyone in the hallways, what do they need you to do with the technology? What's the problem to solve? Right? And so he and I did research together for a while in the lab, and then I ended up drinking the Kool-Aid, realizing that machine learning, artificial intelligence and software, the enabled us to use data with these tools.

00:10:11:21 - 00:10:32:14

Rafael Rosengarten

This was really the future of medicine, right? And so I fully drank the Kool-Aid. I joined the cult, and by way of doing that, I joined the executive team of the company as a fourth co-founder. So we had this Slovenian history. But why is it and special test going forward? Here are some of the key things that that we've gained from maintaining a serious footprint in Slovenia.

00:10:32:14 - 00:11:00:15

Rafael Rosengarten

The first of all is just access to phenomenal human beings. It's an incredibly wonderful group of people, and I don't want to paint an entire country with one brush. But I, you know, everyone I meet there is gracious, humble, thoughtful, brilliant, and really kind of collaborative in nature, but also supremely talented. Right? So it's an incredibly, strong science, technology, engineering and math oriented education in Slovenia.

00:11:00:17 - 00:11:20:16

Rafael Rosengarten

In fact, a lot of what we think of today as modern machine learning has its roots in that part of the world. So there's even a long tradition and history of developing these technologies. And I would argue that two of the top 1 or 2 of the top five most important academic scientists applying AI to biomedicine today are both from Slovenia.

00:11:20:22 - 00:11:44:13

Rafael Rosengarten

And I just had the privilege of working closely with one. And the other came from the same program as my co-founder. So it's it's got this great source of talent. And then there's another characteristic of of the ecosystem there that I think is advantageous, for example, against Silicon Valley or maybe even Boston, rather than having to retrain new talent every year because someone's bouncing to the next shiny thing.

00:11:44:15 - 00:12:04:16

Rafael Rosengarten

We have a deeply committed, deeply loyal, workforce with a lot of longevity. We've got bunches of employees who've been with the company as long as I have. For the whole decade I've been working there, we have deep institutional knowledge, and we punch above our weight because people work so well together, and we can avoid the costs of having to rehire and retrain constantly with this dedicated workforce.

00:12:04:18 - 00:12:11:12

Rafael Rosengarten

So I think a combination of the talent and the loyalty and just the goodness of the people has given us a real advantage.

00:12:11:17 - 00:12:23:11

Christian Soschner

That's good to hear and let that sink in for Europeans. We have smart people in Europe that makes us no doubt CEOs say, let's stay in Europe and not pull everything over to the U.S.

00:12:23:16 - 00:12:47:01

Rafael Rosengarten

Yeah. So and they were structured. I don't think it's so unusual now. I think in the world past Covid, people realized that geography is, is not looked at as a barrier, but but just something to work with. And so, you know, we have our company is registered as a Delaware corporation. Right. This is genius Inc. It's structured in a sort of a sensible, plain vanilla way as possible.

00:12:47:03 - 00:13:08:15

Rafael Rosengarten

You know, we've got a board of directors, etc., but we own a subsidiary in Slovenia, and through that subsidiary we can hire a Slovenian workforce, and then we can also employ us and, and you know, x US Slovenia international workforce through the, the US based entity. And, you know, it's when we talk to investors, when we talk to strategic partners, the structure makes total sense.

00:13:08:15 - 00:13:10:15

Rafael Rosengarten

It's not so atypical.

00:13:10:15 - 00:13:19:21

Christian Soschner

Oh, that's good to hear. We are now 15 minutes in our conversation. And I know that part of the audience will jump off after 15 minutes and come back later to watch the episode.

00:13:19:21 - 00:13:27:14

Christian Soschner

When I think about this part of the audience and say, okay, I've seen the start, and now I make two of my decisions that come back or not.

00:13:27:16 - 00:13:32:15

Christian Soschner

What's the one message that you would give this audience that makes them come back?

00:13:32:15 - 00:13:33:00

Christian Soschner

00:13:33:02 - 00:13:37:07

Rafael Rosengarten

Why should I come back? Well it

00:13:37:07 - 00:13:54:06

Rafael Rosengarten

depends on why they're here. But I think the, that the thing that's going to be most impactful to take away from the whole conversation is how precision medicine is changing, because we are now at this point where we have both data and life sciences that are approaching big data, whereas before it was all small data.

00:13:54:08 - 00:14:02:16

Rafael Rosengarten

And we now have the AI algorithms reaching maturity and an understanding of how to deploy them, in ways that are truly going to change the face of medicine.

00:14:02:18 - 00:14:10:16

Christian Soschner

But that's a good message. When we think about your company, Kenya, Kenya is essentially Kenya and Kenya.

00:14:10:20 - 00:14:16:17

Rafael Rosengarten

In the US, it's the soft G and Slovenia's the hard G. So you choose, I say, like Kenny Ellis.

00:14:16:19 - 00:14:21:00

Christian Soschner

It's code. That's code. I'm from Austria, so basically north of Slovenia.

00:14:21:02 - 00:14:21:20

Rafael Rosengarten

Sure.

00:14:21:22 - 00:14:25:05

Christian Soschner

What's the picture behind your company?

00:14:25:08 - 00:14:56:12

Rafael Rosengarten

Yeah. So the big idea is that all disease. But we focus on cancer. Cancer in general is not one disease, right? In fact, it's not even 30 diseases based on the tissue. It's every patient's disease is unique to that patient. And we now have the tools to understand the uniqueness, to understand the individuality of the disease, and to not only understand that, but then to say, okay, we can predict what's the best possible therapeutic intervention based on the uniqueness of this disease.

00:14:56:14 - 00:15:24:02

Rafael Rosengarten

And so the idea behind our company is to make these tools highly scalable, highly accessible, accurate, informative. We want all of medicine to be precision medicine. And again, we're starting with cancer with oncology. And I think we're we're blazing a path that a lot of others will be able to follow in terms of how to make decision making around everything from developing new drugs to diagnosing and, prescribing new medicines in the clinic.

00:15:24:04 - 00:15:29:21

Rafael Rosengarten

A much more, data driven activity, rather than just following the book or following one one.

00:15:29:21 - 00:15:30:23

Christian Soschner

Scott,

00:15:30:23 - 00:15:38:15

Christian Soschner

what's the toughest problem you're currently working on with your team?

00:15:39:05 - 00:16:01:03

Rafael Rosengarten

I think the toughest thing about building any startup is, frankly, it's the the business side of it, rather than necessarily the product or technology. We've been astonished in some ways about how well the products work, how well the technology's working. It wasn't easy. We've been beating at this for over a decade, 15 years. If you go back to to the R&D in the lab.

00:16:01:05 - 00:16:21:24

Rafael Rosengarten

So it's not an overnight success. But the challenge is really shifting buying decisions, shifting paradigms at the the layer of, well, how do you get a doctor to prescribe a new test? First of all, you have to educate them. They have to, even though the test is available. But how do you change, you know, physician behavior? How do you change, you know, patient reticence to join a clinical trial.

00:16:22:01 - 00:16:40:01

Rafael Rosengarten

So these kinds of kind of ingrained behaviors or and in the US, you know, you have systematic challenges to healthcare. How do you get, you know, easier paths to reimbursement for important tests? Right. These big kind of systemic behavioral or market challenges, I think are the hardest to to crack. And the good news is we don't have to fix that alone.

00:16:40:02 - 00:16:59:04

Rafael Rosengarten

Right? We have a unique, set of tools. We have a unique set of products and a unique selling proposition. But this, this need to to change behaviors and to change, buying patterns in to change government or economic policies is something that the whole industry is pushing towards.

00:16:59:10 - 00:17:15:06

Christian Soschner

It's, good to know when you look at both sides, tech and business. When is the right point for a scientific breakthrough? That's the team starts thinking about the business set.

00:17:15:10 - 00:17:32:05

Rafael Rosengarten

In an ideal world, if I were going to, you know, do my next startup or another startup, you would have the problem that you want to solve figured out well before you have the technology to solve it right. So in a way, you're already proving out the business need by selling solutions before they're built. And sure, to an extent we've done that.

00:17:32:05 - 00:17:53:21

Rafael Rosengarten

But we've also absolutely been guilty of this idea of, well, we've built this really cool tech. You know, we've got this awesome software, this awesome algorithm. Let's go see if anyone wants it. Right. And you really do need to do the market validation first. So I would argue that the business in some ways needs to lead. But you know, it's always a bit of a seahorse or a seesaw seahorse seesaw going back and forth.

00:17:53:21 - 00:18:09:21

Rafael Rosengarten

Right. It's always a bit of a, you know, you first feel a need. And in this case, Genie Alice was born of the needs that I personally felt in the lab, I needed software, I needed better tools to predict experimental outcomes. Right. So we built some of that, and then we went and sold some of that. And we iterate with the customers.

00:18:09:21 - 00:18:21:03

Rafael Rosengarten

And and it has not been a direct like straight as an arrow line from. Here's our vision ten years ago. Here we are today. It's been much more of a windy road than that.

00:18:21:19 - 00:18:27:12

Christian Soschner

When did you decide to turn the the technical idea into a business?

00:18:27:12 - 00:18:46:05

Rafael Rosengarten

Again? This was, you know, I had the good fortune of kind of stepping into something that was already partially baked. Right. So my, my co-founders, I think, had had the passion to be businessmen. Right? I had a passion to keep doing science, but not as an academic laboratory scientist, because I didn't think I'd be very good at that.

00:18:46:07 - 00:19:09:06

Rafael Rosengarten

So my desire was less around building the business and more around showcasing, new capabilities of the technology. Right. And in that way, when I joined the company, I was actually the chief product officer, not the chief executive officer. So I had a bit of a commercial mandate. I had to build something that people wanted to buy. But I would argue that my strength wasn't so much in the business as in daydreaming.

00:19:09:06 - 00:19:27:14

Rafael Rosengarten

Like how how the product should work, how it should look and feel, growing into this idea of, we've really got to have a business led, you know, product roadmap, is something that I've had to learn through mistakes. And I think we're quite good at now. You know, we have a pretty simple thesis. So let me, let me fast forward today.

00:19:27:14 - 00:19:47:15

Rafael Rosengarten

Genius is products. What we sell to the market are AI algorithms that help predict all sorts of things about a patient's response to drugs. So base case, it's a diagnostic algorithm that tells you whether or not a drug is going to work for a given patient, but it does a lot more than that. It helps pharmaceutical companies developing new medicines to figure out how is our drug working.

00:19:47:17 - 00:20:12:12

Rafael Rosengarten

Why is it failing? For whom does it work? For whom does it fail? What are some other strategies should we be combining it with other drugs? Do we be using different medicines entirely? So all of these questions they have around the translational and clinical development of new medicines. And we also work with big diagnostics companies when we have a more mature algorithm to wrap that into an actual clinical test that doctors can order to inform their treatment decisions, that's two pieces of the business.

00:20:12:14 - 00:20:21:14

Rafael Rosengarten

Okay. So what algorithm should we build, right? As my CTO or my CTO, my co-founder told me early on, I can model anything, but you have to ask a good question first,

00:20:21:14 - 00:20:24:11

Rafael Rosengarten

right? So we have to ask ourselves if we're going to build an algorithm

00:20:24:11 - 00:20:31:17

Rafael Rosengarten

and try to work with pharma on drug development, or work with a diagnostics company to get it to market, what algorithm to build.

00:20:31:19 - 00:20:55:13

Rafael Rosengarten

And so we do have a pretty straightforward commercial thesis for that. Yeah. What targets what, you know, genetic molecular targets are a lot of drug companies chasing after because we think they're going to be efficacious. And therefore the competitive landscape of clinical development is quite, quite big. And which drugs are likely, even once approved, to only work for a subset of patients.

00:20:55:13 - 00:21:09:18

Rafael Rosengarten

And therefore they need an algorithm to guide them to the right patients. But these are the two questions we ask. You know, what are the the hot drug development areas. And maybe it's the ones what are the ones that are going to be hot in a couple of years, right. Because you always want to meet the market where it's going, not where it is today.

00:21:09:20 - 00:21:34:11

Rafael Rosengarten

And then the other question is where do patients really need better tools. And therefore we think doctors will use them. And very often it's the answer is the same thing. So drugs like immune checkpoint inhibitors are really, very much aided by the presence of good biomarkers, although we need better ones. Drugs like KRS inhibitors, the Rast family of proteins is responsible for almost a quarter of all cancers.

00:21:34:13 - 00:22:01:08

Rafael Rosengarten

It's the oldest known oncogene. We've known about it for over 40 years, but we only have two approved drugs and those are recent approvals. It's a hyper competitive market to have the next approved cross therapy and to do better than the ones that got there first. So we think we can we can impact that marketplace because we've built, in my opinion, the best and the only commercially available algorithm that can tell you exactly how your patients are going to react to arrest inhibitor, for how long are they going to benefit?

00:22:01:10 - 00:22:21:11

Rafael Rosengarten

If they're not going to benefit what you should do instead? So we try to identify these these competitive drug development situations. You know, it's a bit of a spoiler alert, but we've shared our roadmap publicly. We're also working in the DNA damage response space. There are lots of important drug targets there. And again, the drugs work for some of the people, but not all the time.

00:22:21:13 - 00:22:43:17

Rafael Rosengarten

We're really interested in antibody drug conjugates. It's a new therapeutic modality. There are some approved drugs. Again, they work great for some patients, but not all. And we don't know why we think we can build algorithms that explain all of this and really help bring better medicines to patients. But it's it's now driven largely by. Yeah, what we think the market will, yeah, we'll succeed in the market.

00:22:44:00 - 00:22:54:18

Christian Soschner

That's that's good to know. And then amazing development. When you look at your work in the cancer area in the last ten years, what's the most surprising insights that you gained?

00:22:54:21 - 00:23:09:04

Rafael Rosengarten

Oh, gosh. In most surprising insights in the last ten years?

00:23:09:06 - 00:23:34:02

Rafael Rosengarten

I think that, a key part of my philosophy of leadership and of company building is this idea that if you build a great company, in other words, a place people really want to work that's full of amazing human beings, you know, multidisciplinary, talented that share values but don't share worldviews. Right. Those are two different things. People come with very different perspectives, but aggregate around shared values.

00:23:34:08 - 00:23:57:15

Rafael Rosengarten

You can build a company like that, then a great business will follow. And that's not always obvious. In fact, I would argue that's quite contrary. And to the mercenary approach to building a startup where it's got to be super lean, you know, only ten ex employees working 16 hour days, hustle, hustle, hustle. Building a company where the culture is, is the first thing you work on.

00:23:57:17 - 00:24:01:19

Rafael Rosengarten

Has worked really well for us, and that wasn't obvious in the beginning.

00:24:02:03 - 00:24:07:02

Christian Soschner

How would you describe the culture at your company?

00:24:07:07 - 00:24:26:10

Rafael Rosengarten

But I describe the culture in the terms of the core values that that we each we each promote, that we try to live every day. And, you know, not everyone in the company relates to each of the five values equally. But we all relate to at least one, and we all know all five. And so, they are, people first.

00:24:26:12 - 00:24:53:19

Rafael Rosengarten

That's our North star, right? Putting people first, putting ourselves first, Steve. And sometimes as an example of how to, to put, you know, humans ahead of the non-human, honesty, ownership, constructive and innovation. And, you know, each one of these when we pick, a value or when we think about how to distill down what we believe into a value, it has to be something that you could argue the opposite with a straight face.

00:24:53:21 - 00:25:14:07

Rafael Rosengarten

So no one's going to say, oh, don't put people first. Well, not necessarily. Some people might have an innovation of of, you know, you know, hustle or a value of hustle instead of innovation. People might say, you know, something like, you know, the, the, you know, the least expensive, right? That's not innovative, that's commoditized. So there are ways of taking the opposite position on each of our values.

00:25:14:07 - 00:25:33:17

Rafael Rosengarten

So we've chosen them very deliberately, to aggregate it into a culture where I guess if I had to sum it up, I think everyone in our company, everyone on our team, tries as hard as we do, works as hard as we do, you know, pushes for success because we're doing it for one another, right? We're not we're not in it for ourselves.

00:25:33:17 - 00:25:39:22

Rafael Rosengarten

We're in it for the collective. And so I think that's probably the way that I feel the culture most.

00:25:40:01 - 00:25:50:20

Christian Soschner

It's an amazing answer. I think the most, most of the companies that I experience in deep tech areas or biotech areas, they talk very often about the technology, of course.

00:25:51:17 - 00:26:09:12

Christian Soschner

That they need money and, cash flows to keep growing and keep moving the business forward. But very rarely I to get a conversation about culture and values and beliefs in companies. Why is this so important for you that you focus on culture and values?

00:26:09:15 - 00:26:29:19

Rafael Rosengarten

It's important for a few reasons. But, you know, mainly it's what we do is hard. What we do takes time. It takes patience. And as I mentioned, it's not going to be a straight line. Maybe a better way to think of it. It's not always going to be up. It's going to be up and down. And so I think that focusing on culture, first of all, it makes going to work every day a joy, right?

00:26:29:19 - 00:26:45:16

Rafael Rosengarten

You genuinely look forward to both the ability to do interesting, meaningful work that's mission oriented, but to do it with people you enjoy doing it with. So that's step one. It just makes, you know, it makes the 9 to 5 a lot more fun. And then the other part is it makes us a lot more resilient. All right.

00:26:45:16 - 00:27:20:08

Rafael Rosengarten

So we're resilient to setbacks. We're resilient to adversity. And you know Lord knows in the biotech industry these last four years really since the market burst at the end of 21, it's been really hard for a lot of companies working in the life sciences space. We've been hyper resilient and we punch way above our weight, the sophistication of our technology and our products, the novelty of our products, the ability to solve really hard problems that a lot of people have tried and failed at solving is something that a company our size probably has no business being the best at, but somehow we find a way to to really punch above our weight.

00:27:20:09 - 00:27:23:08

Rafael Rosengarten

And again, I think that all comes from the culture,

00:27:23:08 - 00:27:24:18

Rafael Rosengarten

the continuum.

00:27:24:20 - 00:27:38:05

Christian Soschner

Let's switch a little bit to your company and to your area, your research area that you tackled with your company. When you look at cancer biology today, what's your understanding in the company about how cancer behaves?

00:27:38:05 - 00:28:02:00

Rafael Rosengarten

Well, that's a really great question. And it's you know, it's interesting. There have been a number of thought piece papers recently that attempt to rethink the way that that we understand cancer. I think the most important thing is that cancer is not one thing. It's not monolithic. Right. So historically, cancers have been diagnosed and treated based largely on the part of the body in which they're detected.

00:28:02:05 - 00:28:42:19

Rafael Rosengarten

Right. And in a way, this makes sense because surgical removal of the tumor, cutting it out is still our best or best, treatment. Right. It's the the thing that works, if you can get it all and surgeries done by people who are specialized in the anatomical region at which they surgery, which they cut. So that, okay, it makes sense in a historical context, but what we're learning is that, you know, and what we can appreciate now is cancer is really a disease of of molecules of, you know, mis regulation of the genetic code through its, the way the genetic code is, is realized, through genetic signaling and the RNA,

00:28:42:21 - 00:29:03:02

Rafael Rosengarten

and how it manifests as the molecular and cellular phenotype through proteins, metabolites and ultimately cells. I think it it's perfectly valid, for scientists to focus on any one of those kind of layers of abstraction. Right. So if you think of, you know, a disease model as being an onion, right. You can think of it a core of nucleus.

00:29:03:02 - 00:29:24:20

Rafael Rosengarten

And what's going on? The DNA is the most important. You can go out layers and you can come all the way out to neck. There are companies that think about defining cancer based on molecules that sit on the surface of the cell, right. All of those are valid. The trend we're seeing, though, especially given our ability to measure things and to use AI to analyze them, is really to think of it as an integration of all of those layers.

00:29:24:24 - 00:30:01:05

Rafael Rosengarten

Right? So to have come up with a multimodal, multimodal multimodality model of cancer, now my company has specialized at one of those layers. I would argue we're about the best in the world at understanding the RNA expression layer of cancer. So figuring out how gene expression can be used to, paint a picture of the tumor biology, which is informed by the genetic information which shows up in the RNA and is translated to informs the phenotypic layer.

00:30:01:07 - 00:30:35:04

Rafael Rosengarten

So we like RNA as an input analyte. We think it's clinically robust. We think it's highly informative. It's cost effective. But we recognize there are a lot of good things one could measure and should measure. And so the models that we're building are growing in sophistication in terms of the different layers that inform the models, and what one chooses to actually measure in the clinic, I think, depends a lot on the condition of the patient and what's accessible and what's affordable, so long as the underlying, representation of the disease is, sufficiently well informed, it should work.

00:30:35:09 - 00:30:40:15

Christian Soschner

How can I mention your solution? When do you come in?

00:30:40:17 - 00:30:41:00

Rafael Rosengarten

Yeah.

00:30:41:00 - 00:31:06:24

Rafael Rosengarten

So with most pharma customers, pharma collaborators, we start pretty early. One of the things that makes our approach to building AI models a bit unique is that we are training the model to understand underlying cancer biology. So I want to paint a contrast. The alternative to understanding underlying cancer biology is to say we've got a thousand patients who've all been treated with the drug, and 300 of them did well.

00:31:07:04 - 00:31:25:24

Rafael Rosengarten

And we're going to ask a computer to figure out what those 300 have in common. Right. So that's that's training on outcomes. Perfectly reasonable thing to do. It's how, you know, a lot of machine learning works. But we've built a big model. It's, an AI foundation model. We call it the general supermodel that has learned fundamental cancer biology.

00:31:26:01 - 00:31:51:05

Rafael Rosengarten

And what this allows us to do is to configure products, patient classifier algorithms, translational intelligence algorithms before a drug has ever even been in people. So we start with a lot of pharma companies and we develop these models for their drug. And we can say, well, how do we know if it's working? Well, when you run your your organoid studies, your cell line studies, your mouse studies, we can validate the model.

00:31:51:07 - 00:32:19:22

Rafael Rosengarten

And then when you take your drug into humans, we can do a retrospective analysis on those first in human data and show you the model translates to human biology. It works in the human context. We can retrain the model on human data to make it compound specific, to give our customers a competitive advantage, because their drug is not differentiated by this, this algorithm that can serve as a clinical biomarker and then go all the way to helping build a diagnostic device, if that's in our partners roadmap.

00:32:20:02 - 00:32:44:13

Rafael Rosengarten

But we can start quite early. Now, more and more we're partnering with diagnostics companies as well. And here again, it's a question of aligning the business roadmap. What is the test that we want to put on market. We can configure that algorithm rather quickly and work with our partners to do the relevant validations. The analytical validation, the clinical validation that you need to build a regulated device and then our diagnostic partners will do the work to to actually build the test.

00:32:44:18 - 00:32:59:00

Rafael Rosengarten

And we can share in, in whatever, commercial success there may be. So we have both business models, with pharma, we start quite early in the drug development process and with diagnostics companies, again, it's mostly going to be a market driven question.

00:32:59:00 - 00:33:04:16

Christian Soschner

Where you start basically big pharma early, also in the preclinical phase or late optimization phase.

00:33:04:16 - 00:33:18:06

Rafael Rosengarten

Yeah, I mean, we can jump in anywhere. But you know, typically if you're already in late clinical stage, you've got your roadmap figured out. So, a typical starting point for us would be a drug that's either just pre and or maybe

00:33:18:06 - 00:33:20:07

Rafael Rosengarten

even in phase one first in humans, but where we'll

00:33:20:07 - 00:33:25:14

Rafael Rosengarten

do the initial validations on the translational models and then jump into humans, we can start anywhere.

00:33:25:14 - 00:33:40:16

Rafael Rosengarten

But again it depends a bit on on what the pharma company is looking to get out of it. We've also done a lot of work, frankly, this, you know, retrospective analyzes of later stage trials that either didn't work as well as they were hoping or need to be modified. Or maybe we're looking to say, well, this was successful.

00:33:40:16 - 00:34:07:02

Rafael Rosengarten

What's the next disease we should go to? You know what? We our models are informed by so much biology, right? These these these classifier algorithms we're building have learned a lot of biology. They're quite useful in all sorts of question and answer type activities. So wanting to better understand the mechanism of the drug or maybe why it fails sometimes wanting to understand what a combination strategy would be for patients who need, you know, two medicines to benefit.

00:34:07:04 - 00:34:32:16

Rafael Rosengarten

We have a publication coming out at Asco, the American Society of Clinical Oncology, later this this year, in June, where we looked longitudinally at patients who go on a drug, we can predict exactly how long the drug's going to work. And when it stops working, we can suggest what they should take next. Right. So so developing tools that allow, this kind of insight into the, the patient journey, is, is really, I think, quite unique and transformative.

00:34:32:18 - 00:34:49:16

Christian Soschner

That's good. So yeah, it's amazing the work you're doing. You mentioned a few. It's, let me call it buzzwords. And, I would like to dive deeper into those words for me to understand them better. You mentioned patient classifiers. Can you explain to me what they are and how they flip the biomarker?

00:34:49:17 - 00:34:51:05

Rafael Rosengarten

Yeah, that's

00:34:51:05 - 00:35:07:08

Rafael Rosengarten

a great question. So we struggle a little bit to figure out what's the right word to name our product. You know, what are we selling. And for the last couple of years we've been pushing on this idea of, all right, let's just call them biomarkers. But the problem is that the word biomarker has a lot of existing meanings, right?

00:35:07:08 - 00:35:34:23

Rafael Rosengarten

It's a well-established term of art, and a biomarker could honestly be almost anything when you go to your doctor for an annual checkup and they do your blood pressure, that's a biomarker, right? When they, you know, when they take a blood draw and they measure cholesterol, that's a biomarker, right? Like all of these things are biomarkers. What we're in in cancer and in cancer drug development historically speaking, a biomarker is either thought of as the genetic mutation that we think is causing the cancer.

00:35:35:00 - 00:35:56:03

Rafael Rosengarten

So for example, you have a G 12 C mutation in your class gene. That is your biomarker, which indicates that you get a G 12 C selective KRS inhibitor. It makes total sense. It just doesn't do a very good job of predicting whether the medicine is going to work or not. Or maybe you have overexpression of the, immune checkpoint inhibitor target Pd-l1.

00:35:56:05 - 00:36:21:02

Rafael Rosengarten

Right. So it's a cell surface protein. You have too much of it that indicates that that you should get a checkpoint inhibitor. Again, it makes a lot of sense. You're measuring the amount of target, but it doesn't do a very good job of telling you whether or not a patient is going to benefit, right? In fact, two thirds of 60% of patients who get immune checkpoint inhibitors get sicker, not better, but they still spend a quarter million dollars a year for for that, the right to get the medicine, we can do a lot better.

00:36:21:04 - 00:36:53:02

Rafael Rosengarten

So we we are building biomarker models, but it's a different kind of biomarker, a biomarker reimagined. It's reimagined because it measures the entire transcriptome of the tumor. All of the the data available about both the genetics and the phenotype. And then we distill that through our AI models into a subset of informative biologies biological modules, and those we use to develop machine learning classifier layers that say the patient is going to be drug responsive, non-responsive, or sometimes there are more categories.

00:36:53:02 - 00:37:12:14

Rafael Rosengarten

It could be responsive to the monotherapy. Responsive a combo just depends. We can we can tweak that. So what are what are the patient predictors. What are the patient classifiers. Again they're they're AI models that put a patient into a biomarker bin. But they're much more sophisticated than what we think of is as a traditional biomarker.

00:37:12:14 - 00:37:19:05

Christian Soschner

So it's not just one variable. It's a sum of many variables into the right structure.

00:37:19:08 - 00:37:35:03

Rafael Rosengarten

These are ensemble models and the input features to a, you know, the the input to a given patient classifier is a subset of biological attributes that our foundation model has determined and scored based on the whole transcriptome.

00:37:35:10 - 00:37:43:18

Christian Soschner

But isn't that not only useful for clinical trials, but also for patients, for real patients outside of trials when they are looking for a this.

00:37:44:00 - 00:38:04:14

Rafael Rosengarten

We think it will be so we we as a company have not gone so far as to get a test into the hands of doctors yet, and that's a huge goal of ours. We have some major partnerships, including one we announced at JPMorgan earlier this year, for example, with Tempus. I, we have done work with a number of other, large global diagnostics firms.

00:38:04:16 - 00:38:27:09

Rafael Rosengarten

And the whole goal is to take our scale, our acumen at building these algorithms and partner with a group that has the platform, the distribution and the resources to get those the the algorithms into the clinic. We're hoping to have some, some, you know, have doctors ordering tests based on genius algorithms potentially as early as next year.

00:38:27:09 - 00:38:29:08

Rafael Rosengarten

But, you know, stay tuned.

00:38:29:15 - 00:38:55:05

Christian Soschner

This would be great. Right. So we are basically moving away in the industry from the one track fits all in. And so we don't want to have, the wonder weapon or the controversy that basically eliminates all cancer, cancerous, sense in all patients. We are moving much more in a direction where we want to be precise and say, okay, destruct works with that patient and next track with this patient.

00:38:55:07 - 00:39:20:20

Christian Soschner

Yeah. Now streaming on live stream it. I can imagine that some conspiracy theorists, come up with the idea and say, did you hear what pharma is doing now? The just select the patients, of whom they know it will work in that clinical trial, and then they get it improved and then they give it everybody. Why is precision medicine much better for patients than this, wound up approach where we just try to, eliminate everything.

00:39:20:20 - 00:39:39:11

Rafael Rosengarten

Yeah, it's a good question. So, you know, to be fair, the the inclusion of a, you know, biomarker as a selection criteria in a clinical trial will impact initially, the label of that drug. Right. So you, you can't just say, well, for the clinical trial, we're going to restrict it to this subset and then we give it to everyone.

00:39:39:11 - 00:40:12:08

Rafael Rosengarten

It doesn't actually work that way. But with that said, the the biggest and arguably most groundbreaking cancer drugs, drugs like checkpoint inhibitors were able to be so successful because initially we did in fact use biomarkers to help figure out what are the right diseases, what are the right subsets of patients. And the biomarkers have gotten more sophisticated where we now have what we would think of is, you know, tissue agnostic or pan tumor biomarkers that measure the molecular biology, but it doesn't matter if it's, you know, skin cancer, you know, stomach cancer, whatever.

00:40:12:08 - 00:40:27:19

Rafael Rosengarten

As long as you've got the the indicative biomarker, you can get the drug. And that's a huge shift. Taking this idea that we talked about earlier of thinking of cancer really as a molecular and phenotypic disease, and using that information to direct therapy,

00:40:27:19 - 00:40:33:12

Rafael Rosengarten

you know, the conspiracy theorists can go back to Twitter. They can live our lives if they want.

00:40:33:14 - 00:40:38:00

Rafael Rosengarten

And, you know, go home. But, but I do think that,

00:40:38:00 - 00:41:02:08

Rafael Rosengarten

you know, the idea. Why is precision medicine better? It's because it's where all the stakeholders can come around and agree to something. Right? Drug companies want to have obviously the biggest market for their medicines, but they also want to help patients. The longer a patient can stay on a drug, if we can make cancer a treatable chronic condition, right, by finding the right medicine for that patient, that patient's going to stay on that medicine until you know the rest of their life.

00:41:02:08 - 00:41:27:02

Rafael Rosengarten

That's good business, right? You can create a much bigger and more durable market for yourself if the drug works. Furthermore, for drug companies getting to market fastest, which means avoiding clinical trials, failure and or having the best drug market, so being first or being best in any therapeutic class means the difference between earning about 75% of all the revenue for that drug area versus fighting over the scraps with everyone who comes next second, third, fourth, fifth.

00:41:27:02 - 00:41:50:12

Rafael Rosengarten

Right. So there's a real economic incentive as well for drug companies to run smarter trials to get to market faster, to find the right patient groups. Now, of course, diagnostics companies want to have better tests, right? It's a competitive business. One of the best possible tests. And doctors and patients want the best possible outcomes. So I think there's true alignment and for that matter, hospitals and payers.

00:41:50:12 - 00:42:20:06

Rafael Rosengarten

Right. Insurance companies, national health care systems want better outcomes, right? You don't want to spend $1 million a year on a drug if it doesn't work. So I think that the entire industry can agree that having better biomarkers, biomarkers that are accurate, that are highly informative, is a good goal. One thing that I think makes the genealogy approach different from some other approaches and in a really valuable way, I'm not interested in giving a doctor a biomarker that says, yes, the patient should get drug or no.

00:42:20:06 - 00:42:41:06

Rafael Rosengarten

That patient is completely out of luck. I want to give them a biomarker that says, yes, the patient should get drug, or yes, the patient should get drug, but also a drug B or no, the patient should not get either a or B, they should get C plus D. In other words, I would rather give doctors, the information they need to find the right medicine rather than just a yes or no answer.

00:42:41:09 - 00:42:59:12

Rafael Rosengarten

That leaves patients lacking options. And so this is part of why it's so important to take this approach, where our underlying models have learned so much of the cancer biology, that they can truly direct therapeutic decision making, rather than simply acting as a barrier or a gate to getting treatment.

00:42:59:18 - 00:43:03:18

Christian Soschner

Why do you call your AI model a supermodel?

00:43:03:22 - 00:43:30:20

Rafael Rosengarten

Oh, because I'm terrible at marketing. I so some years ago I think it was in like end of 2020 or early 21, we were working on a pharma collaboration where we had developed a really great predictive algorithm that was built of about 100 genes. 50 of the genes represented the immune biology of the tumor microenvironment. The other 50 represented what we call the angiogenesis biology of the tumor microenvironment.

00:43:30:20 - 00:43:56:13

Rafael Rosengarten

So the vascular system and these two biologies, the immune biology, the antigenic biology are known to antagonize one another. And but and are both important in terms of, you know, tumors receptivity to various kinds of therapies and in fact, very important drug targets sit in the immune compartment of the stroma or the vascular compartments. So checkpoint inhibitors, antiangiogenic like Avastin, so forth, Keytruda, Avastin, big drugs.

00:43:56:15 - 00:44:24:08

Rafael Rosengarten

So we wanted to have a different way of of modeling the interactions in this biology in a way that would let us again predict a whole continuum of therapeutic response and figure out the right medicines. And so once we were successful at this, you know, I was thinking about it. Well, we had two biologists here. And according to, Doug Shanahan, who's an American oncologist working in Switzerland, according to his last paper, there are 14 key biologists, and we've done two of them.

00:44:24:08 - 00:44:43:13

Rafael Rosengarten

So what if we did the other 12? And now we've got a library of 14 model biologists that we can mix and match for any particular cancer, any particular drug. This was the idea, and I sketched this on the back of a sheet of paper on my notepad, to show it to my, my co-founder, my CTO, and ask him, could we build this?

00:44:43:15 - 00:44:52:21

Rafael Rosengarten

And, you know, it was one of those things. I was like, you know, it's like a model, but it'd be more like a supermodel because it would. And it was just a, you know, a name that we, we threw out. But you got to call it something.

00:44:52:21 - 00:44:53:24

Christian Soschner

So

00:44:53:24 - 00:44:58:20

Christian Soschner

if it wants to have to name it again, what would your first choice be this time?

00:44:58:23 - 00:45:21:01

Rafael Rosengarten

Well, now that I see how well it works, I would absolutely call it the genial supermodel. All right, so the supermodel, it's it's incredibly satisfying to see something that was just kind of a fever dream some years ago come to life. But it's really, really cool. Since we announced the existence of this foundation model, back in October of last year at biotech X in Basel.

00:45:21:03 - 00:45:43:05

Rafael Rosengarten

And it started putting it into commercial practice with some of our pharma partners. We are able to build these patient classifiers 15, 20 times faster than we could before. And every time we build one, it gets better, smarter, faster. It's a true flywheel, right? So a flywheel implies this feed forward network effect that every time you turn the pedal on your bike, it goes faster.

00:45:43:05 - 00:46:02:13

Rafael Rosengarten

You turn the pedal on your bike and like a bicycle, we really have two gears here. We've got one gear that's around recognizing new data and placing that data in the landscape of of cancer biology. And every time we see a new data point, whether it's one patient cohort of patients or whatever, that piece of the model gets smarter.

00:46:02:15 - 00:46:26:12

Rafael Rosengarten

And then we've got it ended up not being 14 hallmarks. We've we've now validated about 150 signatures that represent the hallmarks. Every time we we validate a model, the signatures that make up that model, the modules become more confident, become smarter. And so we've got these two, these two cogs that that really let us run fast. And again, every time we build one, we go faster still.

00:46:26:14 - 00:46:36:03

Rafael Rosengarten

So I would I would call it the supermodel again. Although, you know, I'd be happy to take in the comments anyone who wants to, to help us with the naming prize so you can.

00:46:36:06 - 00:47:07:12

Christian Soschner

It's a marketing pro and I could come up with some ideas and then you can choose and select one. But Super Supermodel is basically it's takes I think it's it's simple. It's easy and it sticks. It's a great choice when we think about cancer. I think the treatment of cancer has evolved in the last years. But it doesn't feel to me that we have already cracked everything in cancer, and that people just pop a pill and say, okay, I have cancer, but no worries, I live another 40 years and it's no problem at all.

00:47:07:14 - 00:47:26:17

Christian Soschner

There is still a lot of, let's say, bad cancer treatments going wrong and the patient dies after six, 12 months. And it's it doesn't happen. It doesn't work. Well, what's what's the big problem? That's still not correct. What's your opinion about it? What what what can we do.

00:47:26:17 - 00:47:59:02

Rafael Rosengarten

Yeah. So there kind of two pieces here. The the thing that we know works best is not getting sick at all. Right. So part of the reason why the incidence of cancer, at least in the US and some other places have gone down, but on average is that there are fewer people smoking cigarets today. We had, you know, massive environmental protection laws put in place from the 1970s that started reducing environmental causes of cancer, cleaner air, etc., right.

00:47:59:04 - 00:48:16:11

Rafael Rosengarten

I think there's also in the zeitgeist this idea of, you know, trying to eat and exercise and, you know, so leading a clean life, trying to live healthy, trying to avoid cancer in the first place is still your best bet. The next thing that works best is early detection. And here's an area where there's been some really amazing technological improvements.

00:48:16:17 - 00:48:35:03

Rafael Rosengarten

There's a huge amount of technological promise that's yet to be realized, but I think we will. But the more tools we have to detect cancer early, the better, because if you can catch it early again, you can either kill it with chemo or you can cut it out and it doesn't spread, and you can live a healthy life in remission.

00:48:35:05 - 00:48:55:11

Rafael Rosengarten

You on the end, decide where you need, you know, kind of new drugs, new medicines. It doesn't necessarily mean that patients are going to be very sick by the time they get them. But certainly in clinical trials, that's usually the case, right? New medicines are tested in the toughest proving ground. You could imagine, which are very sick. Patients have already had a lot of different medicines that haven't worked or worked in and stopped working.

00:48:55:13 - 00:49:15:10

Rafael Rosengarten

And, you know, it's just a tall order, you know, human biology is so unbelievably complicated. I once went to a seminar by a physicist who said I quit doing astrophysics because it was too easy. I decided to do biology instead. It was way harder. Right? It's so complicated, and there are so many different ways to think about why.

00:49:15:10 - 00:49:33:09

Rafael Rosengarten

It's it's knowing the right target, finding a molecule, whether it's a small molecule piece of chemistry or a biologic molecule like an antibody or an oligonucleotide finding the right one that is going to inhibit the target, assuming you have the right target. And then there's the delivery challenge. How do you get it to the right place at the right time?

00:49:33:11 - 00:49:52:17

Rafael Rosengarten

And this is part of why genius is so excited about antibody drug conjugates as a new therapeutic modality. This is very much like, you know, it's precision medicine in the sense that it's a guided molecule. The molecule has, a target finder, right? It's got an antibody that's looking for a marker that sits on the surface predominantly of cancer cells.

00:49:52:17 - 00:50:20:15

Rafael Rosengarten

Said this is a tumor. And in that way, we hopefully can kill the cancer without killing the nearby healthy cells. And attached to this by a linker is is a poison. It's a warhead. It's a chemotherapy or some other therapeutic modality. And so this is a an example, but just one example of a new kind of therapy we have in our armament, where we're really trying to target the cancer in a smarter way and deliver the right kind of, drug to that cancer cell, specifically.

00:50:20:17 - 00:50:41:15

Rafael Rosengarten

And, you know, this is, again, it's new. I think these things are going to work really well. But it's not just a question of needing to improve on the drug itself. We need the tools to know who to give them to. Right. And so again, back to what we do and why we do it. We think we've got a system that will give us much more reliable answers about who should get which medicines.

00:50:41:17 - 00:50:46:00

Rafael Rosengarten

Considering that some people are just going to get sick and they're going to need medicines.

00:50:46:00 - 00:51:03:23

Christian Soschner

Yeah, still. But I like I like your approach that you mentioned environmental reasons, for example, not smoking, not drinking alcohol. And it was completely different 20 years ago, 20 years ago. I think the lifestyle here in Europe was more drink, smoked, whatever you want. And when you are sick, you can go to the doctor and then you.

00:51:03:24 - 00:51:05:19

Rafael Rosengarten

Think,

00:51:05:22 - 00:51:12:20

Christian Soschner

Which is a good improvement. The second part that you mentioned is improvement in diagnostics. I think especially early.

00:51:12:23 - 00:51:15:03

Rafael Rosengarten

Yeah. Early detection diagnostics.

00:51:15:05 - 00:51:19:02

Christian Soschner

So we should go regularly to the doctor for a checkup.

00:51:19:04 - 00:51:19:15

Rafael Rosengarten

Yes.

00:51:19:15 - 00:51:36:18

Rafael Rosengarten

But even then, I mean, if you think about your, your wellness. So I, I'm 45 and I just went to have my annual physical a couple months ago. And sure, he does the blood work does this, pokes at some things. But frankly, there's not a single test he's doing that looks for cancer. I had to ask.

00:51:36:18 - 00:51:52:10

Rafael Rosengarten

I raised my head and said, I'm 45 and the guidelines say I should have a colonoscopy this year, right? A regular screening colonoscopy. And at least in the US, the standard is if it's clean, right? If they don't detect anything, then you have 7 to 10 years that you don't have to do it again and then you come back.

00:51:52:12 - 00:52:27:10

Rafael Rosengarten

So I recommend that everyone you know, you should be aware of your age related, you know, diagnostics that you should get. But what I'm excited about are technologies like, grails, test, which has had some success and some, you know, challenges. But the idea that you could have a blood draw and from that regular blood work, you could detect, you know, a large number of potential early stage cancers that, you know, maybe not even a definitive test, but it would be one that says, okay, you should go have a CT scan or an MRI, or you should get an X-ray of this particular part of your body because we detect a signal.

00:52:27:12 - 00:52:38:16

Rafael Rosengarten

And so I think the the early detection molecular tests, are going to be truly transformative. Once you know, they're more routine.

00:52:38:19 - 00:52:43:10

Christian Soschner

Okay. Now, you mentioned that is it really enough to get a blood stroke or.

00:52:43:10 - 00:53:08:08

Rafael Rosengarten

For for some cancers? We I think there was, you know, a bit of a naive hope or maybe just a real, you know, a moonshot ambition that we could detect all cancers from one blood drive. It's not nothing's that easy. Right. But the teams have done a lot of work and we've invested a lot of money. We collectively as an industry, in figuring out what can we detect from sort of the minimally invasive test possible.

00:53:08:10 - 00:53:12:24

Rafael Rosengarten

And, you know, it's not done. It's not solved. But I think those technologies are also really exciting.

00:53:13:10 - 00:53:31:23

Christian Soschner

I mean, my ideal world would be I have an Apple Watch, I have an iPhone. I sit in front of a computer and just get 24 seven monitoring the gadget that I use and get this right. And when something goes wrong in my body, which is ever be a reality, or is it science fiction?

00:53:32:00 - 00:53:32:05

Rafael Rosengarten

I

00:53:32:05 - 00:54:13:22

Rafael Rosengarten

think for most cancers it's probably science fiction. I think for some other diseases it's quite possible. Right. So, digital diagnostics, are really, really neat, especially with things like neurodegenerative disease, even with things like, depression and and certain emotional diseases, the voice recognition tools on your phone already can, can detect all sorts of cool, you know, things that you wouldn't necessarily think could be detected just by having this device, you know, wearable wearables, I think, you know, could have obvious roles to play and things like cardiovascular disease and other diseases where these kind of systems level biomarkers, right.

00:54:13:22 - 00:54:44:07

Rafael Rosengarten

The biomarkers of your pulse, of your blood, of your pulse ox at the skin matter, I think for oncology for better for us we need deeper sampling. That's my guess. But again, we don't expect one drug to cure all does it? All the diseases. It may not be that one diagnostics that's cures all the disease. That's right. You could imagine a urine based diagnostic built into your toilet that helps detect, you know, urinary and gynecological cancers.

00:54:44:09 - 00:55:05:13

Rafael Rosengarten

You know, or other reproductive cancers. You can imagine, one built into your stool test for gastrointestinal cancers, right? They're already great tests. But, you know, companies like Exact Sciences and Garden have tests for, you know, early colorectal detection from, from feces. You know, so it's probably not really a one size fits all, but you could imagine.

00:55:05:15 - 00:55:27:16

Rafael Rosengarten

Yeah. Any analyte that you can get from the body, whether it's saliva, lung aspirate, you know, even just from, from our exhale. Blood, urine, tissues, etc. these could all be used as a suite of kind of devices and say a modern bathroom, to detect cancer. But, yeah, maybe.

00:55:27:23 - 00:55:54:22

Christian Soschner

We started off with, the business perspective on your company. Anything from the business from on the business side to make it as convenient as possible for the patient and customers. The suggestions that you made built it into a normal life. For example, the bathroom. Why not? This would be an amazing way. For once you get to diagnoses where you go and don't need to invest any time for a doctor check in unless you have something serious.

00:55:55:00 - 00:56:16:11

Rafael Rosengarten

Yeah, and you know, there are companies working on this. We as a company landed sort of in the space we are. Again, part of it was recognizing that people are still going to get sick and still need medicines and that, you know, developing new medicines is a major industry that has gross inefficiencies, both in terms of the success rate and the amount of capital spent.

00:56:16:13 - 00:56:40:18

Rafael Rosengarten

That that's just for drug development. And then medicine as an industry of treating patients, has even larger gross inefficiencies in terms of giving people medicines that don't work and, not actually having the efficacy that we want. And so we are tackling something again, that has a huge ethical cost and a huge financial cost. I think we need kind of an all of the above solution space.

00:56:40:20 - 00:56:52:23

Rafael Rosengarten

Yeah, I if our company one day goes out of business, I hope the reason is because we're so good at detecting cancer early that we no longer need medicines for it. But until that day, we're going to try to help patients, you know, find the treatments they need.

00:56:53:01 - 00:57:11:02

Christian Soschner

That's right. Development is a long process. It's expensive and it's really, really expensive. For the last number to write is 5 billion. What do you think? What what change can your solution bring to the industry in terms of cost and drug development? Speed.

00:57:11:05 - 00:57:33:12

Rafael Rosengarten

Yeah. So just to share some numbers with personas and you know, these numbers are sort of widely cited and could be disputed. But plus or minus, you know the the clinical trial failure rate across all trial stages in oncology. So from the moment a medicine or a drug candidate interest trials to getting a market, the success rate is less than 10%, which implies the failure rate is over 90% aggregate.

00:57:33:14 - 00:57:56:08

Rafael Rosengarten

You know that's quite high. Fewer than 10%. It may even be fewer than 5% of new potential cancer. Molecules become medicines. And by the time you get to the clinic, there's typically a billion or more spent on development. It can cost another billion or more to get through all the phase of clinical trials. So, you know, imagine you get a medicine all the way to a big phase three clinical trial, and then you have a surprise failure.

00:57:56:10 - 00:58:18:01

Rafael Rosengarten

That's a very expensive failure. That's 15 years of work, maybe more 2 billion, $2.5 billion of development costs, maybe more. And all the human lives for patients who are counting on that and enrolled in the clinical trial because they they needed something right. So the costs are real. We think that we can decrease the the risk of failure in trials by a lot.

00:58:18:03 - 00:58:36:12

Rafael Rosengarten

There have been, papers, both empirical and theoretical, that show that even the old kind of biomarkers, the simple kinds that I argued at the beginning, are not the ones we should be focused on. Those can improve clinical trial success rates 5 to 12 times in the top five cancers. So you're 5 to 12 times more likely to advance to the next stage of trial.

00:58:36:15 - 00:58:38:21

Rafael Rosengarten

We think we can do better by a lot, right?

00:58:38:21 - 00:58:55:03

Rafael Rosengarten

So at each stage you can think of de-risking the the it's up to the next stage. And if you think about that as a kind of technical economic model, you know, you can you can decrease the discount. You're putting on the potential, you know, net present value of that molecule.

00:58:55:05 - 00:59:25:20

Rafael Rosengarten

And then in terms of efficacy, again, we think that the history of drug development has shown companies are better off having a predictive biomarker than not. And again, you just have to look to the huge successful stories, things like Keytruda. Right. This is the number one checkpoint inhibitor. Tagrisso. You've got all these drugs that became true life saving medicines and hugely important, tools in the physicians armament, because we learned how to target it to the right patients, to an extent.

00:59:26:00 - 00:59:41:22

Rafael Rosengarten

And I think we can keep doing better, but it's clear that it's a winning strategy, both from a practical it's let's get this drug to market, but also from an economic sense and then again, we want to get these tools in the hands of doctors, because we think that's going to make your treatment just that much smarter.

00:59:42:08 - 00:59:55:00

Christian Soschner

Informed decision making. I looked at your LinkedIn profile and you recently posted about the FDA shift in animal models. Can you update the audience on your perspective on that?

00:59:55:00 - 01:00:29:11

Rafael Rosengarten

Yeah, so the FDA recently published a roadmap. And keep in mind it's a roadmap. So it's anticipating being a multi years of phasing out animal models where there's less effective and phasing in patient derived models. The the roadmaps specified things like organoid models. So an organoid is when you take some tissue from a patient and you grow it up in a laboratory setting, but aiming to create a 3D structure that at least in terms of how it responds to a drug in a dish, is going to more closely recapitulate what's happening in the human than just cell lines or again, a mouse.

01:00:29:15 - 01:00:53:02

Rafael Rosengarten

So the kind of drug you can imagine struggles to prove anything. And a mouse would be a human antibody, right? Human antibodies are designed against human antigens. They're intended to be, you know, not just species specific, but sometimes quite individual specific. And so, you know, mouse models, there are extreme workarounds to make mouse models work for these kinds of drugs, but they may not be the most informative.

01:00:53:04 - 01:01:09:23

Rafael Rosengarten

You know, maybe furthermore, kind of globally, we've we've cured cancer in mice. A million times. Right. Even patient derived xenograft where you take a bit of patient tissue, stick it in the mouse and then drug the mouse. We tend those tend to respond to cancer drugs at a very high rate, way higher than you see in clinical trials.

01:01:09:23 - 01:01:31:07

Rafael Rosengarten

Right. So I think the FDA has the right idea, which is we should choose the right model for the drug in the system. I think that everything that's happening at the administration level right now, including at the FDA, is a little bit too blunt. Ax. Right? It's a little too extreme. I don't see us getting rid of mouse models in science.

01:01:31:07 - 01:01:50:12

Rafael Rosengarten

I don't see us getting rid of mouse models in drug development. But I do think there's going to be a greater emphasis on only requiring the models that are truly proved to be predictive or useful for a given question or system. And, you know, this is this should be the case that we you choose the right experimental model for what you're trying to ask.

01:01:50:12 - 01:02:13:11

Rafael Rosengarten

Right? And so through the the journey of drug development, I imagine still using cell lines for certain things. Right. You might use mouse models for, for other kinds of molecules, small molecules or whatever. But maybe for antibodies or biologics, we have something different. Maybe there really is a, a true shift to organoids. And then I would posit that artificial intelligence models have a really big role to play here.

01:02:13:13 - 01:02:34:06

Rafael Rosengarten

This wasn't specifically what the FDA was, was talking about, but the idea that you can train an AI model to truly understand the human disease in a way that you can make predictions that hold up in a clinical setting. Well, we're already doing that. And so I see a future where certain improved AI models and there's going to have to be a very well-defined burden of proof, right?

01:02:34:06 - 01:02:55:11

Rafael Rosengarten

A burden of does this work in the clinic? Is it clinically useful? But if you can meet those standards, I think that AI models could absolutely play a front and center role in, AI and filings in, you know, decision making around getting drugs into the clinic. And, and so that's, that's a future that I think we're contributing to.

01:02:55:13 - 01:03:19:23

Rafael Rosengarten

In addition to my work at geni, Alice, I'm a founding member and a board director of an organization called the Alliance for AI and Health Care. This is and global nonprofit. It's an industry advocacy organization made up of dozens and dozens of corporate members and academic institutes and law firms and individuals who are all united around this vision and this mission of responsible AI.

01:03:20:00 - 01:03:38:20

Rafael Rosengarten

The idea of responsible adoption of AI in health care. And so over the last six plus years as an organization, we've tried to work really closely with the FDA, with the EMA in Europe, with congressional bodies, the National Institutes of Standard and Technology, etc., to help build the framework for how do we know if an AI is working.

01:03:38:22 - 01:03:55:12

Rafael Rosengarten

You know, do we have standards for the training data? Do we have standards for how you report the performance metrics? And we help figure out the right ways to update models or retrain models? Or what if the data is a continuous flow from your Apple Watch? How do you then think about the model getting smarter? These are non-trivial questions right?

01:03:55:13 - 01:04:20:19

Rafael Rosengarten

And we need policy guidance. We need good regulatory laws. So we've always looked at the regulatory bodies as real partners in this, not as an adversarial barrier to adoption, but rather as the necessary partners to figuring out the right way to get these tools in place. And, you know, so I take, you know, the cup half full view of the FDA's, notice the other day saying they're really open to innovation.

01:04:20:21 - 01:04:29:01

Rafael Rosengarten

And let's, let's make sure we as an industry work to do this as a partnership and not in some sort of exploitive loophole way.

01:04:29:07 - 01:04:39:14

Christian Soschner

That it's an interesting movement. I think the stock market reacted very quickly to this announcement. Remember the Red chads River, dropped, a little bit on the market.

01:04:39:14 - 01:04:41:20

Rafael Rosengarten

The Charles River dropped a lot of

01:04:41:20 - 01:04:47:21

Rafael Rosengarten

companies like Recursion Pharma, which are AI native. Starburst.

01:04:47:21 - 01:04:54:14

Rafael Rosengarten

But I'm not smart enough to know I'm not smart enough to play the stock market. And I certainly don't think the stock market is smart enough to, to

01:04:54:14 - 01:04:58:21

Rafael Rosengarten

tell us the real future. But yeah, it's been pretty reactive. It was.

01:04:58:23 - 01:05:21:04

Rafael Rosengarten

Let's put it this way. The FDA announcement came without a lot of preamble. Right. So you see these big market moves when something isn't already priced in. I think Charles River has a very important role to play. It's a very important company. I'm sure that their stocks will rally and their business will succeed. But meanwhile, we see a lot of other companies investing in organoid technologies.

01:05:21:06 - 01:05:48:11

Rafael Rosengarten

And again, a lot of companies like recursion, which recursion has been a real trailblazer building this complete AI stack from tip to tail. But that AI is being trained on data and they're generating the data on these sort of, stage appropriate experimental and human models. Right? So coming at it from the base case of super high throughput cell line drug screens and real world evidence from, you know, humans in the clinic, and you kind of build the AI all the way and meet in the middle.

01:05:48:13 - 01:06:13:20

Rafael Rosengarten

You know, it's a it's a profound, profoundly different approach. And, you know, you can I also host a podcast. It's called Talking Precision Medicine. And if you go back, to our March episode, was an interview with NJ.com. Who's the, the chief commercial and strategy officer at at recursion. And we spoke a lot about how recursion is putting AI in place at every stage of the way, and what that evolution is like.

01:06:13:22 - 01:06:24:16

Rafael Rosengarten

You know, I encourage this audience to go have a listen to that, but, it's a different paradigm. I think big Pharma is quickly playing catch up. And we'll also, you know, have similar tooling up and down the value chain.

01:06:24:20 - 01:06:48:00

Christian Soschner

But we're not there yet. I think this is the point from the FDA announcement. It's a roadmap. It won't happen overnight, probably in five, ten, 15 years. My question to you is when I think back to my time in drug development, when I entered this space in 2006, I have a business background, and but I saw American Clinics was basically a wants to solve three problems toxicity, safety, efficacy.

01:06:48:00 - 01:07:13:16

Christian Soschner

You want to have data on efficacy and safety and scaling up to track ingredient, whatever that is right, and deliver it for clinical stage. When I looked at the numbers at the cost in this area, it's pretty much it's pretty expensive with it where we were really allowed to come and when we then went into clinic in the first company and I talked with, the clinical and the chief medical officer, and asked him, what, what do we get?

01:07:13:16 - 01:07:35:06

Christian Soschner

But why do we need to do this toxicity trials and the efficacy rates over the again, I mean, we already have two months. He said his models don't give us any information as if we should be twisted. And you made your statement in your post that you said that some AI models are already outperforming traditional animal models. So my thought backstage was, why can't we automate this process?

01:07:35:06 - 01:07:42:13

Christian Soschner

I mean, if it gets so little information from animal models, why can't we automate that? We can have it be a reality.

01:07:42:19 - 01:08:14:03

Rafael Rosengarten

Again, I think it's there's not a blanket answer, but yes, I do believe that there will be AI models that have been proved to show, you know, highly correlated or highly predictive, metrics for things like toxicity or dose finding, right, that are going to be used as part of AI and defiance to get drugs into humans, into the clinic that will supersede or obviate the need to go through certain animal models that have not been shown to be predictive, that are done due to old legislation or old habits.

01:08:14:05 - 01:08:40:07

Rafael Rosengarten

This is not going to be true across the board, but it'll be true in some places. And for some drugs. You know, if the FDA is smart and, you know, I think that FDA Commissioner McCarry is a is a very smart person. He's, you know, dedicated his career to trying to improve patient safety. You know, if we keep patient safety as the watchword for the FDA, then that means we don't want to do things that make drugs more dangerous.

01:08:40:07 - 01:09:00:15

Rafael Rosengarten

But we do want to do things that make them more efficient and maybe even safer. Right. I'll give you an example from from cancer right now. Right now, you know, kind of the standard in finding a dose for a cancer drug. And you do this through phase one studies is what's called the maximum tolerated dose. MTD. Right. The FDA has a program called Project Optimist.

01:09:00:15 - 01:09:24:00

Rafael Rosengarten

That's been going for a while, I hope I hope it survives the the administration change where they're trying to think about optimal dose selection rather than maximal dose selection. Right. So instead of just shoving as much medicine in as you can without making the patient sicker, can we find the right amount? That's actually the biologically effective dose. And our contention and something else is working on is we can build AI models that help with that.

01:09:24:02 - 01:09:40:22

Rafael Rosengarten

Right. To, to define a biologically effective dose and what are the right biomarkers that tell you on it for a given patient, maybe that dead that biologically effective dose even varies from patient to patient. So this would be a whole paradigm shift. Right. But it would be better for the patients because again, you're giving them less of.

01:09:40:24 - 01:09:59:23

Rafael Rosengarten

Yeah. Medicine. Medicine is just a well dosed poison. Right. You're giving them less of the poison. Likewise with predictive biomarkers. If we can predict who's going to benefit and you get a menu of options of what the, drug regime should be for those patients, then you can really boost the therapeutic index even of drugs that are less well tolerated.

01:10:00:00 - 01:10:26:20

Rafael Rosengarten

And so this is this is exactly what the kind of stuff we are working on at genealogy. It's can we build these algorithms that solve translational and clinical decision making problems from everything from toxicity and safety to dose finding, but again, predicting therapeutic outcomes, predicting combos, predicting a sequencing so that patients get the best, safest, most efficacious treatments possible.

01:10:26:22 - 01:10:53:04

Rafael Rosengarten

And like I said in my post, some AI models are already outperforming and in some ways, we have a poster, that'll be presented with a pharma partner at the American Association of Cancer Research conference in two weeks. And it's an early kind of data. So it's, you know, take it with a grain of salt. But in this model that we've built to try to understand, you know, which patients are going to respond to which drugs, we're doing translational analysis.

01:10:53:04 - 01:11:20:07

Rafael Rosengarten

So here we're at re analyzing data from some of their pre patient models. Right. Things like organoids and PD access. And we're able to predict response. In in many cases where where those models fail. And so I think that that the AI models that we're building and a lot of our friends in the industry working on other technologies are building are already there in terms of the capacity to outperform traditional translational models.

01:11:20:09 - 01:11:42:09

Rafael Rosengarten

But where we're behind is we haven't built the necessarily the clinical proofs yet, where you can just walk in the door and say, we use this model, it's well known, it's well defined. All the parts are here. The the performance metrics are here. This is what it says. Therefore we are proceeding with our filing, you know, building the you know, the pace of clinical data for any given model or any given device is a long road.

01:11:42:11 - 01:11:47:16

Rafael Rosengarten

It's a road we're on, but we're, you know, the industry as a whole is is not there yet.

01:11:47:18 - 01:11:49:08

Christian Soschner

Yeah.

01:11:49:08 - 01:11:55:05

Christian Soschner

I mean, I've seen this three of fully automated pre-clinical phase with this be this be possible over. Yeah.

01:11:55:05 - 01:11:56:06

Rafael Rosengarten

Well the you know, the

01:11:56:06 - 01:12:26:00

Rafael Rosengarten

person who who talks most publicly about that is my friend Alex Jaramillo from In Silico Medicine. Right. So his company has as there's also AI native they're a drug discovery company, a drug company, drug development company now. And he's been investing very heavily in automated robotics labs in closed loop. You know, can we identify a novel target, identify a novel chemistry, synthesize the chemistry, run all those tests and get it all the way to AI and as fast and automated as possible.

01:12:26:00 - 01:12:44:11

Rafael Rosengarten

And they've probably done the best job of that of anyone or at least he's done the best job of marketing his capabilities of doing that. Is it possible to go from, you know, literally novel target discovery to, to idea first in humans in an automated way? I think Alex would say yes. I don't have a hands on experience.

01:12:44:11 - 01:12:49:00

Rafael Rosengarten

So I'm going to duck the question and say, you know what? I'm not sure.

01:12:49:03 - 01:12:54:01

Christian Soschner

Simulated and automated within one man. If you're applying this, which this would be.

01:12:54:03 - 01:13:14:13

Rafael Rosengarten

Yeah, I, I remain a big believer in human in the loop for a lot of reasons. I think that humans and computers are good at different things. Right. I'm not even going to comment on what robots are good at, but humans and computers are good at different things, and I think that part of what appeals to me about science is the artistic side, the creative side.

01:13:14:13 - 01:13:38:17

Rafael Rosengarten

And I think there's still a very big role for human intuition, for the complexity of humans ability to connect the dots for, for kind of the scientific arts to play a role in this. And so I think that maybe that's why I'm skeptical that it'll ever be fully automated. But it's also I'm not sure that it should be I don't know.

01:13:39:07 - 01:13:48:08

Christian Soschner

Is it really taking the humans out of the process? I always thought it's more, shifting creative work to the humans, but, getting them out of repetitive work.

01:13:48:08 - 01:13:49:11

Rafael Rosengarten

Right. Well, today

01:13:49:11 - 01:14:13:20

Rafael Rosengarten

I would argue that's the case, right. So. And the I would also argue that's the goal that that the goal of AI across health care, whether it's in a drug discovery setting or in a medical practitioner setting or whatever it'd be to give humans more time for human things, right? More time between doctors and patients, more time for empathy, more time to ask questions, more time to think deeply about the patient journey and what the right intervention should be.

01:14:13:22 - 01:14:31:18

Rafael Rosengarten

And likewise, on the drug discovery side, more time to design the right experiments, to choose the models that are truly representative of the question at hand, etc.. So I think that is what's happening and the goal. But I'm pushing back a little bit on this vision of a fully automated process, because then you have to ask, where do the humans fit in the loop?

01:14:31:18 - 01:14:51:12

Rafael Rosengarten

Right. You know, I saw this great meme in reaction to sort of the the rise of the ChatGPT of the world, right? The LMS and it says, you know, like, I don't want an AI, but that's going to make new music and make new art for me so that I have more time to do the dishes. Yeah, give me a robot that's going to wash the dishes so I can go play music and do art.

01:14:51:14 - 01:14:55:21

Rafael Rosengarten

Right. And I think that's got to be the right idea.

01:14:55:24 - 01:15:10:08

Christian Soschner

This is a good point. It reminds me of our preparation, and I read in your material that you sense to me that you almost became a marine biologist or a fine dining chef. Can you can you talk more about this?

01:15:10:11 - 01:15:26:18

Rafael Rosengarten

I've almost spent a lot of things short. So if you look at the books up here, this one is, the professional chef. That's the the textbook from the Culinary Institute of America where I've never gone to school, but I have eaten. And I've also got up there a book on sea turtles and Texas coral reefs and, yeah.

01:15:26:18 - 01:16:01:24

Rafael Rosengarten

So the bookshelf behind me, contains multitudes of lives. I grew up and my, my first scientific passion was, was marine systems. I wouldn't have called it that at age seven. I, like I learned how to snorkel, with my uncle, who was a cell biologist on the south coast of France. He worked in a laboratory called, Italian zoology in the French summer and was a biologist of the old guard who did, you know, really impactful, important cell biology and early molecular biology, but understood that you had to choose the right model.

01:16:02:03 - 01:16:28:09

Rafael Rosengarten

Right. So his model systems for understanding how sperm and egg found each other and how fertilization commenced, where organisms like the sea urchin and jellyfish and and their certain properties of their sperm and egg cells in the biology that made them useful for this. And it turns out, a lot of those kinds of discoveries, you know, and you can extend this to see elegans worms and fruit flies and lots of other organisms underpin everything we know about biomedicine, right?

01:16:28:13 - 01:16:52:02

Rafael Rosengarten

Human as a model, organisms, a relatively, you know, new thing in terms of experimental perturbation. So anyway, I learned to snorkel in his lab on holiday when I was seven, and we went back again when I was ten, and we spent a series of summer vacations, with my family and my cousins and uncles family in Woods Hole in Cape Cod, Massachusetts, where there's also a marine biological laboratory.

01:16:52:04 - 01:17:11:02

Rafael Rosengarten

And I just love that stuff and figured that's what I do. I ended up in college studying evolutionary biology, still kind of inspired by the idea of being a natural historian and going out into the jungles or into the coral reefs to to figure out how natural systems had evolved. And I was starting to think, okay, cool.

01:17:11:02 - 01:17:31:14

Rafael Rosengarten

Can we, can we harness some of these systems? Can we find compounds or chemicals synthesized in nature that has interesting effects and, you know, based on their evolutionary purpose? But somewhere along the way and even younger than seven, I remember cooking with my dad my whole life and love to cook and, at some point the hobby became more of a profession.

01:17:31:14 - 01:17:50:01

Rafael Rosengarten

In college, I would earn money from catering student events, and then I worked in some restaurants and was really on the fence. I love doing both of them a lot, and so I worked in a lab for a while at Yale after college and got a bit bored. And so then I went and worked in a kitchen and back and forth and then I said, listen, I've got to get off this fence.

01:17:50:01 - 01:18:08:05

Rafael Rosengarten

But before I commit to graduate school, which is a very long commitment, let me go see it. Try this cooking thing, because I can I can walk through that door and walk back out if it's not for me. And it was the most fun I'd ever had. I spent, a year in my early 20s cooking at a restaurant that ended up getting two Michelin stars and Upper West Side Manhattan.

01:18:08:07 - 01:18:23:19

Rafael Rosengarten

You know, I was just the cook. I wasn't a chef, but it was still learning really great knife skills, learning how to work with a lot of pans. One of our, one of my cook co cooks there, he would say, if you don't have all eight burners going you're wasting your time. I mean, it was it was intense, right.

01:18:23:21 - 01:18:40:16

Rafael Rosengarten

But I also, you know, I realized that, like, there was a lot of living to do outside of a hot kitchen. And again, thinking of my uncle as as an example, he had this great scientific career. But I also remembered on the summer holidays he would invite all the colleagues from the lab to the house, and we do big, fancy dinner.

01:18:40:16 - 01:18:58:00

Rafael Rosengarten

And you know, he you know, you can you can have cooking as a hobby, as a scientist, but it's very hard to be a cook and have science as a hobby or anything as a hobby. It's an all consuming profession. So I ended up going back into the lab for graduate school, and, I would be lying if I said I'd never look back.

01:18:58:00 - 01:19:00:10

Rafael Rosengarten

I actually think about that decision every day of my life,

01:19:00:10 - 01:19:17:06

Rafael Rosengarten

and I even had an offer the following year, after my first year of graduate school, to go work as a cook. Staffing, a private motor yacht in the Mediterranean for the summer. And, I mean, I'm 25 years old. Like the idea of going to, you know, spend my summer cooking on a yacht in the Mediterranean.

01:19:17:06 - 01:19:26:14

Rafael Rosengarten

Like, who says no to that? I said no to that so I could keep doing science. And I think about that choice every day, you know? But the long life and I still get to cook.

01:19:26:14 - 01:19:27:12

Christian Soschner

So

01:19:27:12 - 01:19:33:11

Christian Soschner

those are sounds amazing. And then, you worked in a michelin star restaurant. This is really great.

01:19:33:11 - 01:19:50:04

Rafael Rosengarten

Well, I have to. I just, you know, this was 20 years ago, and I feel guilty enough that I'm still using that example as, like, you know, something great I did, but, this was before Michelin had come to New York. So we were I will say we were Zagat rated 28 or 29, which was the guidebook of reference back then.

01:19:50:06 - 01:19:57:20

Rafael Rosengarten

But shortly after I left the kitchen, Michelin came to the city and the restaurant won. It stars. So maybe I can take credit. Some credit for that?

01:19:57:20 - 01:20:07:10

Christian Soschner

Absolutely. But what fascinated you and proceeded to pursue, but led you to precision oncology?

01:20:07:20 - 01:20:32:15

Rafael Rosengarten

I'll be honest, it was a bit opportunistic. I mean, again, my my entrance into the entrepreneurial world was being amazed by being inspired by the the technology. And when I say the technology here, part of what I mean is machine learning and artificial intelligence. I had the good fortune of training alongside some people who I consider some of the most brilliant mathematical minds of our generation.

01:20:32:17 - 01:20:51:05

Rafael Rosengarten

A woman named Marina. Is it Nick who's a faculty at Harvard? I would I would put my money on her as the most important academic creator of AI for health care today. Like, just in terms of what her lab produces. She had trained with us on imagery less. Which curvature at Stanford. They're both Slovenian, by the way.

01:20:51:07 - 01:21:11:16

Rafael Rosengarten

And my directement Slovenian mentor, Blanche Zupan. Who's on the faculty at University of Louisiana and his students, you know, getting to train with these guys. I fell in love with the tech coming out of the lab. I wasn't working in a cancer lab. I was working in more of a basic biology lab, looking at issues of self non-self recognition and kind of the evolution of the immune system.

01:21:11:18 - 01:21:32:19

Rafael Rosengarten

But I was a molecular biologist, and I was also inspired by the the ubiquity of our ability to sequence, to sequence the genome, to sequence the transcriptome, the proteome, to be able to actually measure all the molecules and then do something with it with the AI. That was so inspirational. But why precision ontology? Well, again, you need to have a market for the thing you're doing.

01:21:32:21 - 01:22:00:23

Rafael Rosengarten

And if you ask what's the least efficient place to develop drugs, the answer is cancer. And if you say, you know, where is the place where, despite all the investment of time and money to develop new medicines, patients are still failing to thrive. Well, you often will come to cancer again. And so we saw a real opportunity to have an impact and a quick impact on patients, but a big impact on on the overall medical ecosystem, at least at the time, 50% of all pharma R&D went into cancer.

01:22:00:23 - 01:22:27:24

Rafael Rosengarten

But like, again, just a trickle of drugs came out. What if we could make that much more efficient? Would we? Would we be able to make medical prices come down maybe. Right. So it was kind of this market based inspiration of where to point to problem. You know, that was ten years ago now in my mid 40s, everybody I know has been impacted by cancer personally, either a parent or a loved one, a sibling or even themselves.

01:22:28:01 - 01:22:33:04

Rafael Rosengarten

And so the mission has become a lot more human since we've been than working on it.

01:22:33:04 - 01:22:53:23

Christian Soschner

That it's true cancer is a reality, so nobody can deny them, over 50. So, you know, that's what we meant when I think about, cancer oncology, artificial intelligence and machine learning both. How challenging was it in your life to unite these two areas in the last 20 years?

01:22:53:23 - 01:23:08:00

Rafael Rosengarten

I mean, the challenge has been real because, you know, the science is hard to do. All of the science is hard to do. It's hard to build good AI. It's hard to, you know, generate and acquire all the data you need to do the analysis. But I really don't think that genie else has been working in a vacuum.

01:23:08:00 - 01:23:30:10

Rafael Rosengarten

It's been, you know, a common enterprise and we have great partners who've contributed a lot. You know, pharmaceutical companies like, like Debbie, a farm in Lausanne, Switzerland, with whom we work closely. You know, they've got an entire portfolio of strategic investments in companies like ours developing AI and digital technologies to help with cancer diagnostics and a bit of it's selfish, right?

01:23:30:10 - 01:23:55:09

Rafael Rosengarten

Some of us work with their pharma company to help them develop their drugs, but it's also magnanimous in the sense that they're investing in technologies that will help humankind. Companies like Tempus AI, with whom we have a collaboration, you know, they're way bigger than us. They've done a much more public job of bringing next gen sequencing to the masses and amassing a bolus of data that the rest of the world can now come in and use to build these applications.

01:23:55:11 - 01:24:15:01

Rafael Rosengarten

And that's just two examples. One on the pharma side, one on the diagnostic side of partners who have also helped move the needle. So yeah, it's hard and we're not there yet. But I also think there's a it's a real community effort. And you know, we each you know, find a giant stand on the shoulders of and try to try to move the needle as much as we can.

01:24:15:04 - 01:24:28:13

Christian Soschner

Let's talk a little bit about dealmaking and fundraising and show you raised $13 million during, the toughest time for fundraising, I guess 2022 and 2023. How was that for you?

01:24:28:18 - 01:24:48:10

Rafael Rosengarten

Long? It was hard. I don't think it's unique to genealogists, but. But we've had what I would call just monumentally bad timing in terms of, like, when we've got a great idea, when we've got a breakthrough for proof point and we're ready to go to the next phase, the market just seems to conspire against us. So 2021 was a really good year for us.

01:24:48:10 - 01:25:05:00

Rafael Rosengarten

Granted, it was a really good year for a lot of companies in biotech as the market was zooming. But I don't just mean in terms of revenue. I mean technologically, we really solve some hard problems in 21. And we we formalized our vision of the supermodel and we had these proof points with the FDA and with the, you know, in the market with some publications.

01:25:05:06 - 01:25:28:23

Rafael Rosengarten

So it's time to raise money and then the market collapses. Right. So it was 54 weeks from my first pitch in January 2022 to our signature of our, you know, signing the the closing documents on the series A and mid-January 2023. Right. That's a long fundraise. And you know the secret to success? There were three. I think the first one is is deep conviction.

01:25:29:00 - 01:25:47:04

Rafael Rosengarten

I knew that we had solved the hard problem and I knew that if we could get some investment dollars, that we could really blow it out of the water. I, you know, with with the work we have been doing with a company called Angsana, we developed this proof point around our tumor microenvironment classifier. It worked better than anything else.

01:25:47:04 - 01:26:21:23

Rafael Rosengarten

And I knew that we could scale this technology. So just having that conviction is key. Right. And it was a conviction that I had not had previous fundraising campaigns where I was a little less certain about our vision. That's one, two is an incredibly generous network, both of my seed investors and my board members, but also colleagues, you know, really opening their networks to me to make introductions to the right investors so that we could even with our backs against the wall in a tough year, we could be picky and we could be choosy, and we could build a syndicate of smart investors that we wanted not just money, but but smart money.

01:26:22:00 - 01:26:42:14

Rafael Rosengarten

And the third is, is the culture and the team we built. Right. The company, you know, we had to batten down. We had to survive 54 weeks in a bad market to raise this round. And that meant that everyone on the team had to play their part about in the same direction and tighten their belts at times. And we did, and we were successful at it.

01:26:42:16 - 01:26:52:03

Rafael Rosengarten

But, you know, it wasn't easy. But I would maybe ask, is 2025 a better market to raise money? And I'm not sure we'll see how long it takes me to get this round together.

01:26:52:06 - 01:26:54:12

Christian Soschner

Is it ever a good time to raise money?

01:26:54:14 - 01:26:54:20

Rafael Rosengarten

Yeah.

01:26:54:20 - 01:27:00:20

Christian Soschner

I don't know. Oh, you said you mentioned 53 weeks from the first pitch to do something like that.

01:27:00:20 - 01:27:02:13

Rafael Rosengarten

53 or 54 weeks? Yeah.

01:27:02:15 - 01:27:12:13

Christian Soschner

When you count, the time in cash in your bank account, plus preparation up to the first pitch, how long was the time?

01:27:12:16 - 01:27:16:09

Rafael Rosengarten

Oh, when we started pitching, we probably had like three months of runway.

01:27:16:11 - 01:27:18:03

Christian Soschner

Oh, really?

01:27:18:05 - 01:27:23:16

Rafael Rosengarten

Okay. So, I mean, the trick to flying a plane when you're bolting the wings on is don't hit the ground.

01:27:23:18 - 01:27:24:12

Christian Soschner

Yeah.

01:27:24:14 - 01:27:26:07

Rafael Rosengarten

Right. So so, you know,

01:27:26:07 - 01:27:42:19

Rafael Rosengarten

we I mean, we're a commercial organization. We kept selling things and making more revenue. We have a brilliant guy running finance who is very good at budget discipline. You know, we we figured it out. We kept doing business and kind of being just in time. We stayed off the ground. But, you know, it.

01:27:42:21 - 01:27:44:10

Rafael Rosengarten

It's never easy.

01:27:44:13 - 01:27:53:00

Christian Soschner

There are particular techniques and at the same time, you close the deal with a partnership with your farm. You mentioned it before.

01:27:53:00 - 01:28:18:15

Rafael Rosengarten

Yeah. So so these forms are really interesting organization. Again, I think it's an organization with huge foresight and vision. Right. They really believe in precision medicine and pull all of the drugs and development that they work on with precision tools. And I think the reason they've got this vision is because their business model is to in-licensing molecules, either in a pre-clinical stage or maybe phase one, and they get them through phase two, which is the major piece of risk.

01:28:18:17 - 01:28:36:01

Rafael Rosengarten

And then they, you know, re license the molecules now that they're de-risked. Right. And so if you think about what Debbie Farm is trying to succeed at, they're trying to de-risk, you know, new cancer drug candidates to make the medicines. And if you think about what genealogist is trying to do is we're trying to de-risk new cancer drug candidates to become medicines, right.

01:28:36:07 - 01:29:02:16

Rafael Rosengarten

So mission alignment is really good. And we just take different stabs at it. We're working on the patient classifier models. They're working on, on the classical drug development approach and bringing those together is helpful. So Debbie a farm has an innovation fund. They have a venture fund and which I mentioned earlier has a whole portfolio of really excellent and cutting edge companies working on digital solutions to cancer diagnostics and so forth.

01:29:02:18 - 01:29:24:10

Rafael Rosengarten

So we were able to secure W farm as a co-lead investor of our series A, and at the same time, we're negotiating with the pharma side, a collaboration around, one of their assets W0123, which is a, an inhibitor of we one that's a target in the DNA damage response space. And a very promising target, but a tricky one.

01:29:24:12 - 01:29:42:22

Rafael Rosengarten

But these two transactions are separate. So we we earned the investment on the merits of the investment. And we earn the collaboration on the merits of the technology. Now, of course, the two arms of W farm will speak to each other. Right. And in fact, one of the real value ads they give to their portfolio companies is visibility into the pharma's side.

01:29:43:03 - 01:29:56:17

Rafael Rosengarten

And you get a board observer from the pharma side, etc. but they don't do deals just because one group did it, the other one, you know, not every portfolio company gets a collaboration and not every company they collaborate with gets an investment. We had to earn both of those separately.

01:29:57:04 - 01:30:18:12

Christian Soschner

Let's go to the I think the US president will probably say A deal's a deal and I can do everything. And it's pretty much the same. Everything is, is easy. But when you look at both raising your risk and euphoric experience with securing a farmer partner, on the business development side, what's the biggest difference between fundraising for a series A and securing a farm?

01:30:18:12 - 01:30:22:23

Christian Soschner

A partner is a strategic partner.

01:30:23:00 - 01:30:30:23

Rafael Rosengarten

It's a really good question.

01:30:31:03 - 01:30:56:04

Rafael Rosengarten

I mean, I think that in principle, what you're being judged on or is, is a bit different, you know, raising a series a, you have to have some proof points, but the vision is still a really big part of what you're pitching, right? You're trying to sell a future where if you get everything right or even get some of the things right, you know you're going to change the way business is done, you're going to change the way the world works, and it's going to be monumentally valuable.

01:30:56:06 - 01:31:19:22

Rafael Rosengarten

So I think that at the series a stage anyway, you have quite some leeway to to sell a big vision when you're doing a business deal. Yeah. You need to align your vision with where the customer is going. But and they want to know you know what? How does this help me. Right. So one of the biggest pitfalls, I think in startups trying to do business with big companies is we're constantly trying to sell them our technology.

01:31:19:24 - 01:31:41:18

Rafael Rosengarten

Right. Our technology is so cool. Biotechnology. No one wants to buy technology. People want to buy solutions to their problems. Right? So the business development side, it's really the challenge and something I still struggle with every single time I write some piece of marketing collateral. We have to go back a day later and fix it, because I wrote about how cool our technology is and not about what is the customer's problem we solve.

01:31:41:18 - 01:32:04:11

Rafael Rosengarten

Right. So I think that's maybe the biggest difference is in the case of fundraising, you're really selling a future of, you know, a vision of the future that's transformed, transformed because of your your success. And in the case of a business deal, you're really trying to sell the customer a better version of their future. This friend transform because you understand their problems in a unique way.

01:32:04:13 - 01:32:08:17

Rafael Rosengarten

And those are maybe subtle, but but big differences.

01:32:08:20 - 01:32:21:12

Christian Soschner

That will make a fantastic 52nd clip in my opinion. Good understanding of the understanding the customer need. I think we really very often forget this small little. But the most important point?

01:32:21:15 - 01:32:25:16

Rafael Rosengarten

Yeah. And it's embarrassing. You know, we every so often we'll go back and it's like, all

01:32:25:16 - 01:32:39:16

Rafael Rosengarten

right guys, it's time to update the website. And every time we do it, I'm embarrassed because the previous version was too much about us and not enough about our audience. And even knowing that, I can guarantee you that when we update the website, it should be in the next couple of weeks.

01:32:39:18 - 01:32:44:04

Rafael Rosengarten

It's still three months from now going to feel too much about us and not enough about the customer, right?

01:32:44:05 - 01:32:52:03

Christian Soschner

Yeah, and I think this is the most important question. When you talk to a customer, what do you need? What's your problem? Problem? Do you want to solve?

01:32:52:05 - 01:32:52:23

Rafael Rosengarten

Yes.

01:32:53:00 - 01:33:09:03

Christian Soschner

Ken. For the for the young entrepreneurs in the audience, can you walk us through the debris of, till the major steps? How does it work to, engage your former partner with understanding the need and finally closing the deal? I think we need the entrepreneur stream of that.

01:33:09:05 - 01:33:10:05

Rafael Rosengarten

Yeah, I'm. I'm

01:33:10:05 - 01:33:27:18

Rafael Rosengarten

actually not going to use w farm as an example, partly because it's very real and hard to abstract, and partly because it was, you know, it was a long set of conversations and my head of business later, but instead I'll, I'll try to make this more general, and what how we engage with farmer customers more generally.

01:33:27:20 - 01:33:48:02

Rafael Rosengarten

So, again, our products are these machine learning algorithms that are used as translational in clinical intelligence. They're patient classifiers. And the way that we see the world is or at least try to is through the customer lens. A pharma company is developing a drug against a given drug target. A drug has a certain mechanism of action, and that's why it works.

01:33:48:04 - 01:34:14:15

Rafael Rosengarten

So we build products specific to these targets. And I'll give you an example. One of our products is called genealogies cross ID I've already said in this called it that Keras is a major cancer drug target. And there's a whole class of new drugs or KRS inhibitors. Right. So we've built this product for that, that little sliver of the market companies developing drugs against k rest or other rest drugs, rest targets.

01:34:14:17 - 01:34:48:21

Rafael Rosengarten

So here when we engage with the pharma client, we have several big ones, several big pharma and some biotechs working on that with these algorithms. The way we typically start is proof of concept. They need to trust that this algorithm works, right. Why should they believe little old us? So we can do a turnkey evaluation in an under a month, using whatever translational data they're willing to share to show that the base case algorithm, the thing we've already built is sitting on the shelf right here, is going to work to show that it's learned the biology necessary to understand their drug.

01:34:48:23 - 01:35:06:16

Rafael Rosengarten

That's the first thing we do takes a lot of months. Then we work with them to design a proper proof of concept study. What do we need to show you over maybe three months of work, to prove that this is something you need in the clinic to get to that point of conviction, to adopt this as a tool for clinical decision making.

01:35:06:18 - 01:35:26:04

Rafael Rosengarten

And typically these will be retrospective. Studies were first, will retrain the model to be more compound specific and then run retrospective studies on either mouse trial data or first in human phase one data. Or it can be flexible, right? We're quite flexible, but the goal is the same is to get them to a point of conviction that this is the tool they can't live without.

01:35:26:06 - 01:35:53:11

Rafael Rosengarten

And then the next phase of the collaboration is to, commit to to this co-development of the algorithms to support clinical decision making. And here they can license access to the algorithm. We have a whole technology suite that comes with it. And we'll provide the, the, AI enabled expert services to make sure the model is, you know, tailored to their drug, to make it compound specific, to give them a competitive advantage through clinical development.

01:35:53:13 - 01:36:22:17

Rafael Rosengarten

But that's typically how a pharma collaboration will progress. Now, what I'll say is that every pharma we work with has a different process of engagement in terms of where contract team procurement sits, who the champions are, who the budget holders are, etc. so it's it's hard to say, you know, there's a push button way to do this. But to your point earlier question, the key is when you get to meet with those customers or those potential customers, you should do more listening than you do talking.

01:36:22:19 - 01:36:37:18

Rafael Rosengarten

You should do more asking questions of you know what? What are the what are the things you're struggling with? What if you know, here's my favorite question if I gave you a crystal ball and you could ask, look in and find the answer to any of your questions about your drug, what would those what would those questions be?

01:36:37:20 - 01:36:54:05

Rafael Rosengarten

Right. And that's what we're trying to give people. It's not magic like a crystal ball. It's AI, it's science. But if you had a crystal ball to uncover anything about your drug, what's the first thing you'd ask? And what's the second? What's the third? And that's our and that makes our collaboration roadmap.

01:36:54:08 - 01:37:19:02

Christian Soschner

Why why is this so important in in making a deal with an industrial partner to ask a lot of questions instead of this, human notion to start talking and presenting and try to persuade and, try to talk them into buying it. Why is asking questions the better way?

01:37:19:05 - 01:37:47:18

Rafael Rosengarten

They're kind of two answers, right? So when? When your only tools a hammer, everything looks like a nail. So if you're going around just saying, like, let me hit this, like, you know, you're going to end up bruising some thumbs. The it's partly because not only does every pharma company do deals differently, every time a company develops drugs differently, there are real scientific cultures that are specific to organizations and it is presumptuous of us to assume that this tool we've built is the thing you're looking for.

01:37:47:20 - 01:38:07:16

Rafael Rosengarten

For us to think that we understand your pain, it gets back to empathy right? You know, you need to you need to listen and you need to ask not just the first question, but the second and third order questions to really appreciate what someone's going through, whether that's an emotional turmoil or why is developing this drug more challenging than that?

01:38:07:18 - 01:38:27:18

Rafael Rosengarten

And even to companies that are developing competing drugs against the same target may have very different challenges, right? One might be dealing more with a toxicity issue, whereas one might be dealing with a selectivity issue, for example, right. And I would contend that our system of AI, that genius has built our super model can help address all of these questions.

01:38:28:09 - 01:38:43:20

Rafael Rosengarten

But we need to ask the right question, right? We need to make sure that we're addressing the question of its most urgent, most painful, most valuable to solve today. And then we can move to the next one. Or we can do them in parallel. But we first need to, right, you know, to ask the questions and write them down.

01:38:43:23 - 01:39:04:01

Christian Soschner

So it takes a little bit of patience selling pharma till I can imagine it. I mean, the fast way, the ideal way for many people probably would be, here's my tool. It's the best. You need it. Buy it. Okay. We're done. Yeah. How long? How long does it take to secure pharma partner time wise when we talk about.

01:39:04:04 - 01:39:13:09

Rafael Rosengarten

So so the the canonical industry average is like 18 months for an enterprise deal with the pharma company. And a lot can happen in 18

01:39:13:09 - 01:39:27:04

Rafael Rosengarten

months. Here's some of the things that can happen. The drug program that you're trying to work on can get canceled, right. Or be sold. The champion you've cultivated at the farm, a partner that wants to use your product, can retire, get fired, or change jobs, right?

01:39:27:06 - 01:39:44:10

Rafael Rosengarten

The whole team can be reorganized. We had a great call last year with the care team at a big Pharma. Was super eager to get started, and then they went completely silent for three months and we finally got one of them to respond to an email to confided to us that every single person on that call had a new job, and none of them were working on that molecule anymore.

01:39:44:12 - 01:39:53:24

Rafael Rosengarten

This happens. Right? So 18 months is the canonical sales cycle we have right now.

01:39:54:01 - 01:40:25:10

Rafael Rosengarten

Just signed a pilot agreement with the top pharma. We've just completed a pilot agreement with a pilot project with another top pharma, and we have an ongoing, project with a midsize biotech, all for one of our products. Right? Same same set of algorithms for different format and a few others, in the works and we met all of them about 11 months ago, and it took the short the quickest of those was five months to kick off, and the longest was, you know, 11 months to kick off.

01:40:25:12 - 01:40:44:11

Rafael Rosengarten

And I think that's quite fast and it's quite fast because we have a very well defined product that, again, can be aligned with a bunch of different problems or, you know, needs. And that's just for one of our products and we have a bunch of other products, but I'm using this as a kind of a trying to make a generalizable example.

01:40:44:14 - 01:41:05:18

Christian Soschner

And I said, you mentioned that you needed 53 weeks, to secure funding for your company, 18 months, the canonical time until closing. If it takes a lot of patience before you talked about it has a culture in Silicon Valley, and they also believe you have sort of a culture in your company where you push things forward and make things happen internally.

01:41:05:20 - 01:41:15:12

Christian Soschner

Where did you learn patience? In dealing with farm accounts, then what's your secret?

01:41:15:15 - 01:41:21:19

Rafael Rosengarten

I don't know that I have patience in dealing in pharma. I'm going to, you know, challenge the premise here.

01:41:21:19 - 01:41:39:02

Rafael Rosengarten

I'm not sure that I have patience. Where did I learn patience? Stick with it a little bit differently. So. And you know, one thing about Chainalysis, we actually I don't think we have a hustle culture. I think we have a, it's a bit of a work hard, play hard culture, but it's a very balanced culture.

01:41:39:02 - 01:41:58:13

Rafael Rosengarten

We have the ability to to see the long game. And this is what's important. You know, working 80 hours, 100 hours a week is is nonsense in my opinion. Sometimes. Okay. Sometimes there's a sprint. You have a deadline. You can't help it. We have a a program in our company called Flex Fridays. We don't schedule internal meetings on Fridays.

01:41:58:15 - 01:42:27:18

Rafael Rosengarten

You're encouraged to not schedule any meetings on Fridays. You're encouraged to do deep work on whatever you want to work on. And there's a lot of research now showing that four day workweeks are more productive than five day workweeks. There's a lot of research, I'm thinking here of, like, Cal Newport's book Deep Work, showing that giving your employees not only permission, but encourage them to find large stretches in their calendar where they're not disturbed and allowing them to use that time for whatever problems they're trying to solve gets you to the solution.

01:42:27:20 - 01:42:50:02

Rafael Rosengarten

And here's the other thing. I don't care if my employees use their Fridays to come to the office to do deep work, to stay at home, to do deep work, or if they decide that they need to go climb a mountain or run a marathon that day in order to clear their head and think through something. You know, we've got a team of scientists, and in my experience, biologists have a huge stack of papers on their desk and they're just waiting for a long flight.

01:42:50:02 - 01:43:09:04

Rafael Rosengarten

So they can actually have time to read. And computer programmers have a huge, you know, pile of new computer code libraries or new tools that they're trying to find time to try out, but they just don't have time. If you give people a day where you say, go do what you need to do. I don't care what you do.

01:43:09:06 - 01:43:28:10

Rafael Rosengarten

You need to do what you need to do. You end up way more productive. And this is something we put in place two years ago, and it turns out it works like a charm. It's not a day off. It's a day to maximize your well-being and your efficiency however you see fit. So this is not a direct answer of how do we learn patience.

01:43:28:10 - 01:43:49:18

Rafael Rosengarten

But I think that part of the genius culture is to be very process oriented, right? Yes, we have goals. We have a quarterly OKR system for quarterly goal planning. We have, you know, weekly tasks. We have three strategy like we set goals, we run hard at stuff. But we also know that we've got a ten year vision right.

01:43:49:20 - 01:44:13:23

Rafael Rosengarten

And you know that keeping some perspective on how long this takes is important. I think this has shown through my career, my, you know, my adventures underwater, my adventures in kitchens, my adventures in molecular biology and transitioning to running a business. You know, none of this was intentional. It's not like I set a goal for myself by the time I'm 45, you know, I want to be running this company and doing this thing.

01:44:14:00 - 01:44:32:14

Rafael Rosengarten

If you figure out a process, if you figure out a journey that's rewarding, you know, if you find that balance where you can be really productive at work, you can have impactful, meaningful work, but have time for your family, time for your friends, encourage hobbies. Then the patience comes.

01:44:32:14 - 01:44:42:01

Christian Soschner

You mentioned in the preparation material, it's all about the journey, about the destination. Where did this mindset come from?

01:44:42:01 - 01:45:05:24

Rafael Rosengarten

Well, oh, I just, you know, I enjoy doing stuff, I enjoy learning. I think I, you know, maybe it comes from if you're not super goal oriented, it's harder to be disappointed. Right. But part of that is just like, you know, genuinely enjoying the process. I think it's it's very much related to taking so much joy from from learning things.

01:45:06:01 - 01:45:10:01

Rafael Rosengarten

Right. When you get to the end of something, you know, what fun is that? Right?

01:45:10:01 - 01:45:17:06

Rafael Rosengarten

So I love the title of your podcast, so What? Remind the audience what this is called, right? This series.

01:45:17:08 - 01:45:18:12

Christian Soschner

Beginner's mind, beginner's mind.

01:45:18:16 - 01:45:22:13

Rafael Rosengarten

The Beginner's Mind. Yeah. So my wife and I have a challenge to each other.

01:45:22:13 - 01:45:34:06

Rafael Rosengarten

Every year we try to come up with something for each of us. We come up with something that we're going to be a beginner at, something totally new that we suck at, that we're going to try and we're going to try to learn and we're going to enjoy doing it.

01:45:34:08 - 01:45:41:02

Rafael Rosengarten

I've got a ton of hobbies that I really stink at, but man, being a beginner is so much fun, right?

01:45:41:05 - 01:46:02:06

Christian Soschner

That's true. That's true. You run a company and the company is very awesome. All right. But focus. Focus on one thing and become the best in this area. And then you contradicted in your private life and say, no, I want to be a beginner in many things and I know disciplinary. What's the benefit of this multidisciplinary approach in a focused business world?

01:46:02:08 - 01:46:05:02

Rafael Rosengarten

Yeah, I think

01:46:05:02 - 01:46:25:11

Rafael Rosengarten

that is genius. Has one real Achilles heel. So we're not great at focus. And that's probably my fault, right? We do try to do a lot of things, but the way that I excuse it, the way that I, I kind of make it okay is we do have a very clearly articulated vision, right? We went through a strategy setting exercise just before closing the series.

01:46:25:15 - 01:46:52:00

Rafael Rosengarten

So in December 2022, where we drafted a three year strategic roadmap with quite some detail, and I go back and look at that every quarter as we set our quarterly goals. And sure enough, we're right on track. Right? So we have had we've had had a Northstar. We've had a plan. But as it turns out, this field is so damn complicated that all these things that might feel like distractions or side quests turn out to be, and maybe this is not by accident.

01:46:52:02 - 01:47:15:10

Rafael Rosengarten

Essential learnings, essential pieces of the puzzle. All right. You know, we need to get really good at this thing in order to support this other thing. And maybe it wasn't obvious at first, but there was a reason why we're doing it. So, you know, having that kind of Northstar that clearly articulated vision aligning with the company, we have we have, all hands strategy retreats every six months, the whole team together, the leadership team first and then the whole team.

01:47:15:12 - 01:47:33:23

Rafael Rosengarten

You know, these are crucial points in time where we we make sure everyone remembers how their piece rolls up into the next piece and how the functional teams roll up into the whole and, you know, so, so, that's not to say we're trying not to, you know, we try to be nimble. We react to realities on the ground, we react to changes.

01:47:33:23 - 01:47:37:15

Rafael Rosengarten

But we we think we know where we're going.

01:47:37:15 - 01:47:55:05

Christian Soschner

You mentioned side quests. I mean, this happens very often in company states. New things pop up and, some might distract from the mission, others might benefit the mission. What's your decision making criteria? How do you decide what's important for your mission and whatnot?

01:47:55:13 - 01:48:15:10

Rafael Rosengarten

Yeah. I mean, at the end of the day, you know, everyone, there's resource constraints on everyone, right? And venture backed startups, are case studies and resource constraints. You're constrained in money, you're constrained in time. You're constrained in effort. Right. So we have any number of ways. So some are very empirical and granular when we think about software development.

01:48:15:10 - 01:48:34:04

Rafael Rosengarten

Right. But building the tools, we have a, scoring system that determines how much effort something is going to take and how much impact it can have and so forth. So it's straight out of a product management playbook to come up with a way of prioritizing what we work on based on this score, where the numerator is the impact and the denominator is the effort.

01:48:34:04 - 01:48:56:20

Rafael Rosengarten

And right, you're trying to, you know, maximize this ratio, all the way up to much more amorphous things like crap. That's a cool new publication technology. Can we poke it that and see if it's useful? One of the really big successes in the last year for us has been building an internship program, where we run a lot of the more far fetched ideas we had.

01:48:56:21 - 01:49:15:11

Rafael Rosengarten

The side quest off to, to, you know, young talent with the right amount of supervision, but without too much investment that it drains our resources. And we've actually made some really great breakthroughs through this program where we say, oh, that's got promising. All right, let's pull it into the actual R&D team now and make it, make it work.

01:49:15:13 - 01:49:33:17

Rafael Rosengarten

But we don't have a hard and fast, you know, decision tree. But it does take discipline. Some of it is just, you know, having the right conversations with the right people, you know, both at the C-suite level, on the managerial level and even at the ground level to get feedback like, is this working? Is it worth it?

01:49:33:19 - 01:49:41:16

Rafael Rosengarten

When do we decide to cut bait or when do we decide to double down? Yeah, there's there's some art here in addition to the science.

01:49:41:19 - 01:49:59:19

Christian Soschner

Yeah. That's true, that's true. This is the the tension between focus and innovation. Know innovation. And, the focus doesn't matter. On the other hand, just innovation. It just, being distracted and multidisciplinary doesn't lead anybody to here. Finding the right balance is key to success. In the end of the day.

01:49:59:21 - 01:50:00:24

Rafael Rosengarten

Absolutely.

01:50:01:01 - 01:50:17:02

Christian Soschner

Yeah. I see that we are. The time flies when you have fun. And, it's really great talking to you, but I just checked the clock, and, we are on that for the two, two hour mark. What would you like to focus on in the last part of our conversation?

01:50:17:04 - 01:50:17:09

Rafael Rosengarten

Well,

01:50:17:09 - 01:50:34:20

Rafael Rosengarten

first of all, if you're still watching the live stream and you're not my mom, send me a note. I want to know who you are. That you found this so exciting. We can have a call. What do I want to focus on? I mean, you tell me you know your listenership better. What do you think would be most interesting to to double down or to bring up at this point in the conversation?

01:50:34:22 - 01:50:43:14

Christian Soschner

Future predictions? I really like the okay at the end of the conversation to look into a crystal ball and, think about how the future looks like. It looks like.

01:50:43:16 - 01:50:44:17

Rafael Rosengarten

Yeah, future prediction.

01:50:44:17 - 01:51:08:06

Rafael Rosengarten

So I firmly believe that medicine will continue to become more precise. And it's going to happen in a lot of ways. The actual design of molecules will be informed by human biology in ways that the medicines themselves are more precise for some kinds of medicines. I think we're going to be able to take a readout from a patient and make an end of one medicine for that patient, right.

01:51:08:06 - 01:51:26:15

Rafael Rosengarten

These are, you know, things like cell and gene therapies that are really tailored to those patients. And then more generally, I think we're going to have an armament of off the shelf medicines with the tools to direct them to the right people. So I really do believe that all of medicine will be precision medicine, not just in cancer, but in every aspect of medicine.

01:51:26:17 - 01:51:53:15

Rafael Rosengarten

I think it's inevitable that artificial intelligence is going to be in everything we do. I am honestly surprised at how fast that's changed. Right. Because it was slow and it was slow and was slow and people were like, yeah, right. Artificial intelligence, nice buzzword. And then all of a sudden it's literally flipping everywhere, right? There was this inflection point with, I think, you know, the open AI, sort of, movement into the public domain.

01:51:53:17 - 01:52:12:05

Rafael Rosengarten

So granted that I think it's going to be everywhere. I think you're going to see a lot of consolidation. Right? So some big players harnessing some of the tools, I think you're going to see, you know, the things that work rising up and a lot of stuff that doesn't fall by the wayside, but a lot of experimentation, we're not going to know what works at first.

01:52:12:07 - 01:52:35:17

Rafael Rosengarten

And my hope is that the stuff that really helps improve health care, improve well-being, well care and secure, those are two different things, right? Improve people, you know, help them live healthy lives, help them detect disease early, help and find the right treatments. This is going to be very much, a human computer partnership. Going forward.

01:52:35:24 - 01:52:53:01

Christian Soschner

When we think about the next ten years and assume that your company is wildly successful in the industry and everything that you are working on really works out well and changes the healthcare industry. How does the industry look like after ten years?

01:52:53:01 - 01:53:20:18

Rafael Rosengarten

Yeah. If genius does our job right and our collaborators, you know, make the contributions we expect every single cancer drug is going to have, a suite of algorithms that tell you, how safe is this molecule? How much should you give and who should you give it to? Every single cancer drug. And I think we can use the learnings from drugs that are in development and drugs that are in the market to help us find the next best targets.

01:53:20:18 - 01:53:50:05

Rafael Rosengarten

Right. So new targets, and new chemistry, to save lives. So I really think that we're going to be part of the glue that takes this long road of drug development for target discovery and compound discovery to patients on market, and makes it from one straight line to a circle. Right. We want to close the loop between the human experience and the discovery of the next generation of medicine and, make, you know, all of that as precise and efficacious and efficient as possible.

01:53:50:12 - 01:54:08:04

Christian Soschner

But does it mean that artificial intelligence in drug development sets, every drug will have, a suite of AI tools connected to it, then AI is not so much a tool. It's more a companion to the drug at the end of the day. So it evolves with the drug.

01:54:08:06 - 01:54:08:11

Rafael Rosengarten

I

01:54:08:11 - 01:54:10:18

Rafael Rosengarten

think that's right. I think I think

01:54:10:18 - 01:54:15:18

Rafael Rosengarten

every drug is going to have kind of an information companion that helps it on its journey

01:54:15:18 - 01:54:17:19

Rafael Rosengarten

to move faster, right?

01:54:17:19 - 01:54:27:10

Rafael Rosengarten

You know, again, already in silico medicine has a proof of concept where they move the drug from a claims from target discovery and compound design to AI. You know, first in human in 18 months.

01:54:27:10 - 01:54:55:19

Rafael Rosengarten

That's lightning fast. Now the question is, is that drug going to work better. And we're waiting to see the readout. Right. But imagine you that's going to speed up even more. Imagine every investigational drug to get to a decision point. And I any decision point in 12 months. Right. And then now imagine that they're actually better and they're better either because they're design better or against better targets or going to be all the above because they've got algorithms that help dose them properly and help find the right patient groups.

01:54:55:21 - 01:55:07:24

Rafael Rosengarten

This could be a huge it could be, you know, 12 months or preclinical. And can we shrink development to a 4 or 5 year process with a, you know, 75% success rate? That's a different world.

01:55:09:04 - 01:55:19:17

Christian Soschner

Yeah, I hope I hope it works out in that way. More precision in the drug development process, faster timelines and better drugs, faster on the market at a cheaper price.

01:55:19:19 - 01:55:21:07

Rafael Rosengarten

All of that. Yeah.

01:55:21:09 - 01:55:42:06

Christian Soschner

I think it's I think it's possible. I mean, 20 years ago when I thought about the heart problem, I always learned it in commercial school. You either can push the price down or the quality and both at the same time. It's not possible. And then we had this outlier. Companies like Apple, Amazon, Microsoft, Tesla, they really cracked everything on all ends.

01:55:42:06 - 01:55:47:11

Christian Soschner

So why not track development? We are still lagging behind in that area in the.

01:55:47:13 - 01:55:47:23

Rafael Rosengarten

But I

01:55:47:23 - 01:55:57:09

Rafael Rosengarten

think it's because the failures take so long and cost so much that the successes have to cover those costs. Right. So again, if we can make it faster, make them better

01:55:57:09 - 01:56:02:19

Rafael Rosengarten

so there's less failure, then I think you can have, you know, a better product for cheaper.

01:56:02:21 - 01:56:11:21

Christian Soschner

What's the biggest bottleneck that you see in the industry currently that holds that holds us back to achieve this goal?

01:56:11:23 - 01:56:12:10

Rafael Rosengarten

In my

01:56:12:10 - 01:56:35:07

Rafael Rosengarten

opinion, it's a misalignment at the payer level. I think that we are still Pennywise pound foolish in so many cases, but here I'll speak most about the US health care system. It's takes a huge lift to get reimbursement for a diagnostic test when we know if that test works, it's going to save you a lot of time and money.

01:56:35:07 - 01:57:02:00

Rafael Rosengarten

On the health care side. But a lot of, you know, the paradigm is our insurance will pay for an expensive drug that may or may not work, but not for a relatively cheap test that'll let you know. So the piece that I'd love to see change is a dramatic simplification of the way health care is paid for, especially in the US, and a real appetite, top down appetite from payers and providers to see outcomes as the key driver of how we do things right.

01:57:02:00 - 01:57:06:24

Rafael Rosengarten

We need to deliver health care value. We need to deliver outcomes not just activities.

01:57:07:01 - 01:57:08:00

Christian Soschner

Yeah, that's important.

01:57:08:00 - 01:57:16:14

Christian Soschner

I feel at the end of the call, what's one big ask for the audience? What are you looking for with your company?

01:57:16:16 - 01:57:16:24

Rafael Rosengarten

One

01:57:16:24 - 01:57:22:01

Rafael Rosengarten

big ask. I'll give you three. We are going. We are always looking for for

01:57:22:01 - 01:57:41:21

Rafael Rosengarten

pharma to collaborate with. So if you've been listening or if you tune in and you're developing new medicines, even if it's not an ontology, I'd love to have a conversation about how we could team up and do a better job. If you are an investor in the space and you you share our vision and and find our track compelling, you know, we're always we're always raising money, but we're actively raising money.

01:57:41:21 - 01:57:58:10

Rafael Rosengarten

Now, I'd love to have a conversation, you know, and if this is the kind of organization you think you'd like to work with or work for, you know, get in touch. We're actually not hiring today, but we're always recruiting. We're always building relationships. Because at the end of the day, when we do add to the team, it's about culture fit.

01:57:58:10 - 01:58:09:09

Rafael Rosengarten

It's about values matching. And so we want to get to know you. Reach out. I'm especially responsive on LinkedIn. You can leave some comments in the show when we post the show. I'd love to have a conversation.

01:58:09:12 - 01:58:12:02

Christian Soschner

So LinkedIn is the best way to connect with you?

01:58:12:02 - 01:58:13:02

Rafael Rosengarten

Yeah.

01:58:13:04 - 01:58:23:18

Christian Soschner

It's good to know. Yeah. Okay. Looking back at the conversation, what's the biggest message you hope sticks with the audience after today, after listening to the episode?

01:58:23:20 - 01:58:47:19

Rafael Rosengarten

Oh, gosh. And this is changing the face of medicine. It's a long road, and it's, it's a very heavy, boulder to push up that hill. So let's do it together. I think that if we all, you know, can can find that kind of patience and enjoyment in the journey, you know, it's going to make it a lot more, a lot more bearable to do it together now.

01:58:47:23 - 01:59:10:02

Christian Soschner

Well, thank you very much for this fantastic conversation. I enjoyed every single minute of this two hours. Thank you. Question. I think the most important message for me was that we really should work hard on making the drug development process more efficient for the betterment of the patient, and stick it out. At the end of the day, less sick patients means a better society and less pain for families.

01:59:10:04 - 01:59:12:08

Rafael Rosengarten

100%. I couldn't agree more.

01:59:12:10 - 01:59:18:03

Christian Soschner

And I said, I wish you a happy rest of the week and let's stay in touch.

01:59:18:05 - 01:59:21:00

Rafael Rosengarten

Thank you Christian, this was a real pleasure. Thanks for the opportunity.

01:59:21:02 - 01:59:24:02

Christian Soschner

Thank you very much for being on the show. Have a great day. Bye.

01:59:24:03 - 01:59:24:14

Rafael Rosengarten

Okay.

01:59:24:14 - 01:59:29:12

Christian Soschner

Precision doesn't just belong in labs. It belongs in how we treat people.

01:59:29:12 - 01:59:38:11

Christian Soschner

Today, profit Rose garden took us deep into the heart of biotech's next leap, where RNA, artificial intelligence and empathy converge.

01:59:38:13 - 01:59:46:17

Christian Soschner

We explored a world where every track has an information companion, and every patient gets the treatment that fits them.

01:59:46:19 - 02:00:04:17

Christian Soschner

Health care still wastes billions in drugs that fail not because the science isn't there, but because the systems aren't ready. Ruffin's message is clear. Ask better questions. Build with purpose. Solve real problems. That's just technical puzzles.

02:00:04:22 - 02:00:09:11

Christian Soschner

If this conversation shifted how you think. Don't let it stop at you.

02:00:09:11 - 02:00:18:20

Christian Soschner

Follow the show. Leave a review. Share it with someone who still believes medicine should be personal, data driven and human at its core.

02:00:18:20 - 02:00:22:22

Christian Soschner

every follow. Every share brings this vision to more minds.

02:00:22:24 - 02:00:31:10

Christian Soschner

It brings in guests like Raff and deepens the conversation and helps us build the community committed to reshaping medicine, not just improving it.

02:00:31:12 - 02:00:39:17

Christian Soschner

The future of medicine isn't cold or mechanical. It's thoughtful. It's tailored. It's collaborative. And it starts with people like you

02:00:39:17 - 02:00:48:01

Christian Soschner

curious enough to listen, bold enough to act. Until next time. Stay curious. Stay tuned and hit subscribe.