Rafael Rosengarten - Part 3: Synthetic Biology Meets AI in Precision Oncology

 Intro - 00:00:01: Welcome to The Biotech Startups Podcast by Excedr. Join us as we speak with first-time founders, serial entrepreneurs, and experienced investors about the challenges and triumphs of running a biotech startup from pre-seed to IPO with your host, Jon Chee. In our last episode, we talked about Rafael Rosengarten's graduate studies at Yale and his postdoctoral work at the Joint BioEnergy Institute, where a mix of collaboration and access to cutting-edge technology sparked his passion for synthetic biology. If you missed it, be sure to go back and listen to part two. In part three, Rafael takes us through his move from the Joint BioEnergy Institute to Baylor College of Medicine, where personal commitments and a vibrant medical community informed his next moves. We'll hear about his work in synthetic biology, a transformative collaboration that bridged biology and machine learning, and how these experiences led to the founding of Genialis. Rafael candidly shares the challenges of pivoting the company's focus, ultimately positioning Genialis as a leader in predictive biomarker development for precision oncology.

Rafael - 00:01:22: My girlfriend became a fiance. Her work took her. She was doing a rotational program like every six, seven months. She actually went to Venezuela for her work. So I stayed in the Bay Area. But we kind of made a deal. It's like, if you can get back to the U.S., I'll just come find you. And so that was Houston, right? All oil pipelines lead back to Houston. And so she ended up back in Houston. And I drove back out here when we got married. And so I figured I'd just figure it out here. Houston, not everyone knows this, has the world's largest medical center. It's, I don't know how many institutions now, but when I came here, it was like 68 or 69, shoulder to shoulder. A friend of mine once described it as Vegas for hospitals. It's nuts. I mean, it employs over 150,000 people just in the medical center alone. And so if you have training in biomedicine, it's not hard to find a job here. And I was able to apply for and get a postdoc fairly readily at Baylor College of Medicine, which is a historically phenomenal institution, right? Like really, really storied human molecular genetics department in the lab of Gad Shaulsky and Adam Kuspa, who worked on yet another weird fritter, an organism called Dictyostelium, the cellulose slime mold, doing self non-self recognition. So I basically left all the synthetic biology stuff behind, at least in terms of the biofuels, went back to the problem space that I'd finished my graduate career on, but in yet another organism. I still wanted to do the tools and I love the synthetic biology. So, and they kind of hired me for that. So I wanted to like, just innovate, like, can we do better tools, better, you know, can we do this science faster, better? And in fact, I had the best result. I postdoc there for four years. I had my best experimental result like three months in. Was able to achieve something with molecular cloning of dictostelium DNA, which is like 80% AT-rich, which most people won't understand what that means. But the composition of the genome is highly skewed towards one set of chemistries over the other, and it makes it fragile. So it's hard to work with. And I figured out how to work with it. No one else had been able to do that. Never came close to being that successful in the lab again for three more three and a half more years, yeah, you kind of shoot your shot. 

Jon - 00:03:28: Yeah, yeah.

Rafael - 00:03:29: But at any rate, I was able to get this postdoc with some, again, wonderful mentors. I've always, I don't know if I'm good at picking them or they're good at picking me, but really wonderful people to work for. Again, same characteristics, completely committed to the career trajectories of their trainees, stable in their own work, no egos at all involved, very open-minded to let you set your own course, right, and excited to help you do that. And I put my head down and I did that, but I knew, and this was kind of the weird part, like, this isn't going to be my career. This isn't going to be my forever job. And, you know, I had to kind of figure out what's next. But in the meantime, you know, my wife, who was very supportive of all of this. She was on a pretty awesome trajectory in her job. And so, you know, I was comfortable saying, listen, I'm going to keep doing science and I'm going to keep doing it in a way that remains geographically flexible. So, like, you know, if you want if we want to go overseas for your work, I'll figure it out. Right. And so I had her blessing and support, but also. You know, the stability, financial stability, you know, life stability, family stability, to continue to be a dilettante and open-minded and not destination focused, right? Like I was just kind of along for the ride.

Jon - 00:04:43: Very cool. And were there any cultural and lab cultural, personally, professionally, when comparing Baylor to JBEI, like in that experience, or were they very actually quite similar? You just happened to find another JBEI?

Rafael - 00:04:56: No, not even that different. Like I don't even know how to begin to compare them. The Baylor College of Medicine was a much more traditional, like biomedical research. Place with proper departments and grown-up supervision and HR policies. But also, at least in some ways, more resource constrained. There were very well-heeled HHMI labs, Howard Hughes Medical Institute labs, where I imagine the budgets were just kind of off the wall and you could do whatever. But it wasn't the same. You weren't playing with government paper money.

Jon - 00:05:26: Interesting. When you mentioned that kind of like finding your hit in the first three months, I always imagine this like you're like, all right, now I need to do like my song. It's like a musician and their sophomore album. They're like, crap. How the hell do I how do I just be what I just released just now? And you're just like, yeah, pressure is high and you're just like, it's intense.

Rafael - 00:05:44: Yeah, I kept doing science and part of what I had released was a tool. So I got to keep collaborating with other people in the lab. What happened at Baylor that was so instrumental in my life was actually, it was a funding thing. So again, I was working in this weird, it's not that it's a non-model organism system, it is a model, but it's not a traditional biomedical model. And so funding was always going to be a titer. And so I needed to get a fellowship, right? And my mentor proposed to me that I apply for this NIH fellowship through the National Library of Medicine for biomedical informatics. And what was special about this fellowship, there were a few things. One is it was administered through an umbrella organization called the Gulf Coast Consortium, which was a partnership of six member organizations in Houston. Baylor College of Medicine was one. I want to say University of Houston, Rice, probably UT Health. So it was this multi-organization umbrella. And it required that you have a primary mentor sponsor, but a co-mentor from a different discipline. And the overall thrust was to pair. Bench scientists with computationalists and vice versa. So if you were a lab scientist, your co-mentor would be a bioinformatician or a data scientist or whatever. And you would figure out a project that used both sets of skills. And I was a life scientist, so I had to find people to work with on the computational side. So that was a big piece of it. At the same time, my academic lab at Baylor had a longstanding collaboration with a data science team, computer science group in Ljubljana, Slovenia at the University of Ljubljana. So probably most Americans couldn't point to it on a map. I'll give you the hint. It's at the very top of the Adriatic, nestled between Italy and Croatia, just south of Austria, right? It turns out that some of the most important concepts in machine learning and advanced mathematics grew up in that part of the world. And as it turns out, there was a brilliant young faculty member at University of Ljubljana who had done a master's at University of Houston and somehow gotten a faculty position, that he kept at Baylor College of Medicine. And he would come over for a few months every year, and he would bring students. And so I had the good fortune of being at this lab at Baylor when Professor Blaž Zupan came over from Slovenia and he brought with him a woman named Marinka Zitnik. Marinka may or may not be the smartest mathematician of her generation. In the world. I just mean in the world. Yeah. And I don't mean to exaggerate. She's just next level brilliant. She's currently on the faculty at Harvard. She had done her PhD in Lubyana, postdoc at Stanford. And she keeps just racking up awards and publishing the most incredible stuff at Harvard. This is 2011, right? So they came over to do work in our lab and they actually came for a year. And the idea was they had invented all these new technologies that the machine learning techniques and mathematics that predate what we now know as like large language models or autoencoders and all these major machine learning AI concepts that you can't walk down the street without tripping over now, both in the consumer realm and in the life sciences. But this was 2011, right? Most people didn't know about this. And so all of us in the lab at the bench got to team up with them to think, all right, how can we design some in silico experiments to make some predictions that we can then test in the lab and create a nice little feedback loop of making predictions, testing lab, making predictions in advance, both of our science. And so I got to work every day, one-on-one with Blage and with Marinka on what we now know is like, you know, is at the cutting edge, the absolute forefront of machine learning. Just to give you a little bit of like point in time, Recursion Pharma, which is arguably the most famous of the tech biopharma, it opened its doors in 2013. So this is two years before that. Insilico Medicine opened its doors, if I'm not mistaken, in 2013, two years before that. Benevolent AI, 2013. So this was early, early days of machine learning and life sciences, right? So I had this fellowship or I got this fellowship. I had these collaborators. And I was like, all right, I'm going to put all this together, right? I'm going to make this my postdoc. There was also a cool technology coming out of the Slovenian group where they had built a prototype or an MVP of this lab software for omics analysis. So I'm still doing high-throughput sequencing, still doing the data analysis. But here's what we would call nowadays a code-free solution, point-and-click, super interactive software for playing with the data. And so as a bench biologist who really shouldn't be allowed to code, this was brilliant for me. I could design experiments, do this. And so I was just immersed in this world of software and data science. And it blew my mind. I mean, I drank the Kool-Aid and I went back for seconds and thirds and fourths, right? The fellowship required that I do some interdisciplinary training. So the fellowship administrator who I always say this, I owe her everything in my career, but God, at the time, it felt like torture. She made me go sit and audit a statistical machine learning class at Rice, which has a very strong applied mathematics department. I hadn't taken a mathematics class since freshman year of college. Right. Man, it opened my eyes. But what it allowed me to do is have conversations, right? So I didn't have to learn how to do machine learning. I had to learn how to talk to people who knew how to do machine learning, right? So I'm not giving this period of my life enough credit in the narrative, but in terms of all the events that took place. But one of the events was another one of Blasch's students came over. His name was Miha Štajdohar. He had just finished his PhD in artificial intelligence. And had started this little company with two other guys in Slovenia called Genialis, and they were going to license some technology from Blas's lab. And do something with it. That's all they knew. And so Blage got the company sort of early access to Baylor. I don't remember all the terms because I wasn't really involved, but basically two, Miha and one other early employee, a bioinformatician, could come over to Baylor. And Baylor would hire them as postdocs to build software that did some bioinformatics, did some AI, kind of combined all this for labs at Baylor. And so I was customer zero for this. I was like the earliest customer for some of these tools. And again, drinking the Kool-Aid, like full-throated support of the stuff. It was awesome. I knew I wanted to work on this. I knew I wanted to take this machine learning and figure out problems to solve with it. And Miha and I went rock climbing and camping one weekend up in hill country outside of Austin. And he proposed to me in the tent. He said, why don't you come join us, right? So one year into his two-year stint in Houston. Was already kind of at the end of the postdoc. I left my postdoc, joined the company by way of becoming a founder because the company wanted a U.S. footprint. They wanted a U.S. headquarters. They wanted someone, all three of the previous, the other founders were from other domains. They wanted someone who understood biomedicine and life sciences. For some reason, they recruited a washed up marine biologist line cook who didn't know anything about biomedicine or entrepreneurship. Like I had no business sense, no commercial sense, never a day in my life have I worked in a lab that solved an actual medical problem. But God, I believed in the technology. 

Jon - 00:12:59: I can see kind of this like recurring kind of theme of, well, the first one is kind of like, I think it ties back to like, just like following your curiosity, but also you seem to have been on, I've been using the term ground floor of these like early waves. Like you're at Yale during like the next gen sequencing kind of like boom, then you were at the ground floor at JBEI for synthetic biology. And then now you're kind of at the ground floor of AI/ML. And I've always tried to figure out like, can you manufacture this? Or is there a way to manufacture serendipity? And I think it's kind of like, at least what it seems to me, that it comes back to that curiosity. And also just like fun, adventurous stuff. Because like, if you follow that, like, and you try to tap into that fun adventure, you tend to find yourself in places like this.

Rafael - 00:14:00: I think you're exactly right. I mean, the truth is, if I had been two years shifted in any direction, I probably still would have been at the forefront of something big and important. And this is, I think, how science moves. The funny thing is, at each of these stages, when you would read the scientific history, it felt like all the important problems were solved already. But of course, none of them were, right? It's nonsense, because there are always important problems. I had nothing to do with CRISPR. I've actually never done CRISPR with my own two hands. I was out of the lab and doing in silico stuff before it was commonplace. So I missed CRISPR as an innovation to actually use in my own research. So I didn't hit all of them. But if I had been two years later, I would have been there for the early days of CRISPR. Or if I had been two years earlier, it might have been something different. But I think this is why being okay with not having a plan and being comfortable just chasing your passion is a pretty good survival skill. It's a pretty good approach to these things. If you have faith and some amount of security, things will work out.

Jon - 00:15:03: Absolutely. And I can say from my experience too, it's just like, uncertainty can definitely be anxiety inducing. That's for sure. And it's maybe it's a bit, you know. Overly simplistic where I, or just maybe trite or where I'm like, it'll all work itself out because sometimes it doesn't. And so I don't mean to overly simplify it, but at least for me, when I have kind of tapped into that element of just like going into the unknown, I'm usually pleasantly surprised with what I find. 

Rafael - 00:15:33: Yeah, I mean, it's not lost on me. This is a little bit of like comes from a position of privilege, right? I think you and I both have the first couple of tiers of our Maslow's hierarchy of whatever pretty well taken care of. So we can bask in the upper levels and just really go for self-actualization. But even people who maybe need it to work out for the sake, for economic reasons or whatever, who are willing to kind of trust that it will. I think diving into the process is usually better than fixating on a goal, you know, on an endpoint.

Jon - 00:16:03: Absolutely. And so now we're kind of like Genialis is like now a thing. You are now kind of the starting point of the U.S.. Kind of footbridge. Can you talk a little bit about or more about Genialis, like the driving force for Genialis, the mission and focus and more about you and your founding team?

Rafael - 00:16:23: Yeah, you know, it's not a simple story. And there are different versions of it, depending on the audience. The company has a lot of different origins. So again, there's this trio of guys in Slovenia who kind of started a company because they wanted to do a company. First, they didn't really know which technology they were going to use. They certainly didn't know what problem they were going to solve. They just knew they wanted to do entrepreneurship. But I kind of respect that in a way. They wanted to get into the process. And they had a good intuition for where to look for the technology, right? The AI wave was coming. And they had a sense that life sciences, biotech, biomedicine had a lot of unsolved problems. So all three of those were good intuitions. When I joined up in October 2015, I like to talk about that as like the company's still in stealth. I mean, it had a website. It existed in Slovenia, but it didn't exist in the U.S. yet. And we kind of kept it in stealth here because we wanted to see if we could go after non-dilutive funding. So the way we actually did it is I started a Genialis Inc., in Texas that actually was unaffiliated with a Slovenian one except for some tech transfer agreements. Because I needed to be able to go after NIH money. So we couldn't have 75% of the cap table be Slovenian, right?

Jon - 00:17:32: Yeah, yeah. 

Rafael - 00:17:33: So we kept this kind of weird parallel thing for about a year. We applied for an NIH, it was one of the SVIRs. I don't even remember which. And the idea was to take this machine learning that I've been doing with Marinka Zitnik, not with my co-founder, but with this other Slovenian scientist, which at this point was a published and open source technology. So to take that and start building AI models that would really learn basic cancer biology, it would learn fundamental cancer biology through the relationship of lots of otherwise unrelated pieces of data. So this is going to become important when we fast forward to today. The original vision, my original vision for what we were going to do was to build integrated data models of all of cancer biology from, it wasn't big data, it was loads of small data, right? All of the small data we get our hands on, right? Because at the time, there was really no such thing as big data in biology. At the time, nothing was that high throughput. And then apply that to intractable cancer domains like glioblastoma and pediatric medial blastoma. And others in order to be able to figure out new targets, but also subset patient populations for the right therapies and stuff. This was the original idea. We submitted that for funding, came very close. We got a very good score, missed the funding cutoff by a little bit. And because we'd missed the funding cutoff by a little bit, we had to figure out what the business was going to be. But at the time, we also had the software prototype. And we were able to sell a single enterprise license to that software for enough money because the company was small and we were paying ourselves next to nothing. In fact, my wife's only thing was, yes, you can do it, but don't earn less than you did as a postdoc, right? So I was basically earning my postdoc salary plus $1.

Jon - 00:19:15: Yeah, yeah, yeah.

Rafael - 00:19:17: So that enterprise software deal made us think, well, look, there seems to be a commercial market for the software. Why don't we do that? And I was brought in as a co-founder, but the head of product, chief product officer, because, again, of my co-founders had this delusion that I somehow knew what we should be making. And so the founding CEO, who's still a good friend of mine, although he's not with the company anymore, he had this vision of building a big scalable SaaS company. Now, there was another trend in the industry at the time. And Andreessen Horowitz had recently entered the life science investing market. And Andreessen Horowitz, this juggernaut Silicon Valley fund, they had made their name and they're not investing in SaaS companies and had this mantra, software is going to eat the world. And they were like, we're going to take this software model and apply it to biotech and change how things are done. So there are all these companies starting up at that time who tried to build a SaaS business. Only one of them that I can think of actually did well with that. The rest of us tried, built beautiful software, but could not make the business model work for various reasons. And there are loads of us, like us, Invisigenix, Cyclica, lots and lots of others who started out in that space and then had to pivot into a more discovery mode. We had to build things that had more intrinsic value closer to discovery, whether those are drugs or AI models or whatever. We can get back to that. But yeah, it really didn't work, but we tried. So we were building this SaaS business around the software. We kind of put the machine learning on the back burner because you have to have focus as a startup, right? Strategy is not what you do, it's what you don't do. And so we decided, all right, we're going to put our heads down. We had some early customers for the software. It was beautiful software, really worked well. We have a brilliant designer. She still works for the company and runs a department around communication. And so we were trying to build this kind of like no-code, low-code informatics software that would allow people to do machine learning. But also had a really robust coding engine for the bioinformatician wonks. We built it. We launched partnerships with Roche, with Thermo Fisher. We had some big splashes, got some big name customers, but could not make the unit economics work. My co-founder, CEO, colleague managed to raise a proper institutional deed round. Now, this would have been pre-seed in today's parlance. It was like two and a half million, a little less, almost nothing compared to the seed rounds you've seen in the last five years. But enough for us to do a proper commercial launch of the software. Grow the team's headcount and prove that we were going to run out of money and couldn't sell the sucker. And so that was the first couple of years of the company. First two and a half years, three years, loads of fun, tons of learnings. Again, I didn't know anything about business. When he told me he wanted me to be the chief product officer, I was like, absolutely, I'm all for it. We got off the phone and I literally went and Googled, what is a chief product officer? I'm not even kidding. I spent the next week reading all of the product blogs from Silicon Valley. What does it mean? There are some great books on this. Books like Hooked and Delighted and books about how you actually build products for customers. How do you find out what a problem to solve and solve it? Great training, but our product, people loved it. It still exists. We still use it. We still sell it. We just couldn't get the economics to work. The SaaS business model wasn't the right business model for what we were doing.

Jon - 00:22:37: Interesting. And it's, I had a similar experience too, when I was first starting Excedr, and this was like when like Marc Andreessen was still blogging, but like Pmarca, and like, uh, you know, the Union Square Ventures to the, all of the blogs out of there my co-founder, he's like, go learn how to do sales. And I remember, I was like, I was a bench scientist, I was like, I Googled, how to do sales? And they're just like, I'm just like, how do I do this? And then proceeded to just learn, just like trial by fire on how to do it. And I think too, like even though, this first iteration or this first V1 of Genialis, and the SaaS model, wasn't what you ended up sticking to, exactly what you said, these experiences are all accretive, it's like they all add to your general repertoire and make you a more formidable entrepreneur, um, as long as you can like again, kind of what we're talking about like get through that wall. Or like make the pivot, and so, when you were there, can you talk a little bit about that experience of doing the pivot?

Rafael - 00:23:40: Yeah. So, you know, I'm going to flash forward a little bit. Now we're in 2018, right? So we raised our financing at the end of 2017. I consider that sort of coming out of stealth, right? We had a big launch. We grew the company way too big. And when it was clear that we were running out of money. You know, the leadership team got together and, and We're like, we need to pivot. We need to figure out something else to do. And so the founding CEO had the courage and the self-awareness to be like, I don't know what to do next. And because I was the only one of the four of us who had a PhD in the field we were trying to sell into, they kind of nominated me to be the CEO to go out and just talk to the market, right? In their own language, find out what problems we could solve. Now, we had done two things really well. We had built really high-quality software that did solve a real problem. It was built to aggregate, process, visualize, and manage next-gen sequencing data, which was now starting to reach the high-throughput scale that you needed for machine learning. And we knew from the early days that the reason we built software in the first place is, if you wanted to do the AI bit, you had to have the data. You had to get the data in one place. And we just got sort of, I'll say, sidetracked because we thought we could sell the software as the business rather than doing the second part. So in the summer of 2018, I took over as the CEO. The other thing we'd done really well is we built a brilliant team, like really, really high-quality engineers, scientists. Mostly in Slovenia, which at the time had sort of a salary arbitrage opportunity. It was less expensive, but that was changing too as more global opportunities were coming about. So we were a pretty big team. We were like 30 people that summer. You know, with dwindling funds. And so the first thing I did as CEO, more or less, is I cut headcount by 50%, which was awful. Like I had to fire a lot of my friends. And the next thing I did is I told the remaining 15 people in the room, I'm cutting your salaries in half. Who's with me? And everybody stayed. And so we went back to salad days. We went back to sweat equity days, right? And, you know, it was tough. Like I went and I said, you know, I remember I went over to Slovenia and I said, listen. I don't know what we're going to do exactly, but we're going to go find out. Here are the things we've built that work really well. Here's what I think are the coming opportunities. And we're going to wander into this dark woods. And if you follow me, we'll come out the other side. And everyone did. And... I don't remember the exact sequencing of this, but around Christmas of 2018. So we had been kind of taking a lot of meetings with a lot of biotechs and figuring out what problems we could solve for them. I had kind of figured it out. There was a confluence of some things happening. So first of all, biotech investments were starting to heat up. End of 2018, this is when interest rates were pretty low for a while. Capital was starting to feel kind of free. Stock markets were all right. Venture capital was starting to come into biotech a little faster than it had before. The other thing is that immune checkpoint inhibitors had proved their worth. And the fight was fierce, right? But Keytruda was starting to emerge as the winner. But everyone and their brother wanted to develop a combination therapy with Keytruda or one of the other ones. And so I talked to a lot of biotech execs. And what I learned was these drugs had a ton of money coming out for clinical development. These drugs did not actually work for that many people. The people they work for, they work really well, but they only work for a few people. And we could use our technology. We could use a combination of the software, but also get back into this AI bit that I loved so much. And we could do the thing that I had written into that SBIR grant three and a half years earlier, which was, let's build some models and figure out who's going to benefit from these drugs. And sure enough, we got a couple small service contracts to build predictive models to kind of reanalyze some retrospective clinical trial data and see if we could predict post hoc who would have done well on these drugs. And so I wrote a note to the team at the end of 2018. I said, this is what we're going to be for the next phase. We are going to be the predictive biomarker company for immune oncology. And it needs US badly. And so we started getting service revenue. And by the middle of 2019, we were breakeven. And we could start paying ourselves real salaries again. And we were able to just grow in revenue. By the end of 2019, we had forged a really good collaboration with a company called Oncerna Therapeutics. It doesn't exist anymore. It didn't survive the biotech winter. And it's been recapitalized as another entity. They were a therapeutics company with some drugs that played in the tumor microenvironment related to the immune oncology. And a vision for building a predictive biomarker system using RNA sequencing data, which is what we had as a company specialized in it by that point. And so by the end of 2019, a year after I drafted this memo that we were going to be the biomarker company for immune oncology. We were the RNA biomarker company, not necessarily for just IO, but we were really good at building predictive models using transcriptomics data, using RNA-seq data. And we had solved some of the really hard... Incredibly boring and unflashy problems around using gene expression data in a clinical context. And loads of people have been playing with gene expression signatures for various uses. There are a lot of risk scores out there, activation scores out there. There are very few clinical devices that use it. And at the time, there were none, really. And we cracked the nut on some very, again, hard but boring problems that I don't think anyone else had solved. And that was sort of the aha moment, right? We did it within this collaboration with Anxerna. I will forever be grateful to the CEO, Laura Benjamin, who took a bet on US and became a good friend of mine in the process. I remember meeting up with her at J.P. Morgan 2020, and we shook hands on a year-long pact to bring their biomarker to market with their drug. And by the end of 2020, we had an FDA-accepted biomarker for use in their pivotal clinical trial. She went on to license it to three global diagnostics companies for a CDx development and other development. And we were working with lots of other biotechs to do retrospective studies, to potentially use it for other drugs, not just uncarnish drugs. This thing worked like gangbusters. Genialis since been a co-author on studies of over 15,000 patients samples in retrospective studies across 12 different solitudes for, I think we're up to 8 different drugs. Maybe 9 drugs, half approved the other half not approved. And in every case, this biomarker model works to enrich for patient response to drugs, hitting the microenvironment, from everything from checkpoint inhibitors to VEGF and back again. And we really learned how to do what we do now through the course of that work. But maybe more to the point, we built the technology engine that we have today, you know, through a lot of blood, sweat and tears. But that was the discovery process of what are we going to be when we grow up.

Outro - 00:30:32: Thanks for joining us on this episode of The Biotech Startups Podcast, featuring Rafael Rosengarten. Be sure to tune in for part four as Rafael reflects on scaling Genialis, navigating the biotech industry during the COVID-19 pandemic, and pioneering cancer diagnostics with AI and machine learning. If you enjoyed this episode, subscribe, leave a review, and share it with your friends. See you next time. The Biotech Startups Podcast is produced by Excedr. Don't want to miss an episode? Search for The Biotech Startups Podcast wherever you get your podcasts and click subscribe. Excedr provides research labs with equipment leases on founder-friendly terms to support paths to exceptional outcomes. To learn more, visit our website, www.excedr.com. On behalf of the team here at Excedr, thanks for listening. The Biotech Startups Podcast provides general insights into the life science sector through the experiences of its guests. The use of information on this podcast or materials linked from the podcast is at the user's own risk. The views expressed by the participants are their own and are not the views of Excedr or sponsors. No reference to any product, service or company in the podcast is an endorsement by Excedr or its guests.