#2 Noor Shaker of GTN on Disrupting Drug Discovery
By Nasos Papadopoulos, EF Head of Content
Noor Shaker is the Co-Founder of GTN, a company that uses quantum physics and machine learning to search the huge space of potential drug molecules and discover drugs that were previously hidden from view. This means GTN can discover better drugs, more quickly, at half the cost.
Noor and her co-founder Vid met on EF8 and after combining their expertise in machine learning and quantum physics respectively to found GTN, they’ve made rapid progress, recently closing a big funding round.
Noor was Professor in Machine Learning and AI at Aalborg University in Copenhagen before she joined EF and has ten years of academic experience in the field, producing more than 50 papers with 1000 citations and an entire book on generative networks.
In this conversation we dive into Noor’s personal story, everything she’s learned about finding product market fit and funding in the process of building GTN and what she’s learned about herself along the way.
Building the Future of Drug Discovery
Reducing the Drug Discovery Cycle from 5 Years to a Month
Stay a Professor or Become a Founder?
If You Build It, Will They Come?
How to Choose the Right Vertical as a Technologist
Nasos: 02:09 Noor, welcome to the show.
Noor : 02:09 Thank you. Thank you for having me.
Nasos: 02:12 It's a pleasure to have you on. I'm really looking forward to diving into your story and finding out what you've learned along this entrepreneurial journey of building GTN.
Nasos: 02:20 To kick us off, give me a quick 30 second intro into what you guys do at GTN.
Noor : 02:28 Okay. What we are doing is we are building a novel technology to not just make drug discovery faster, but also make it more efficient and open up a new space of potential drugs.
Noor : 02:40 We're doing that by combining ideas from quantum physics, machine learning, and apply that to biochemistry. Which is a unique interdisciplinary thing.
Noor : 02:53 The way we're doing it is basically advancing the representation of chemical compounds and then building a whole stack of machine learning tools that can learn from this representation and to predict chemical activities and then sample new, novel chemical structures.
Noor : 03:09 We hope that by doing this, drug discovery won't be anymore limited to what people know about drugs. But it will be more open to the whole space, like astronomically huge a space of potential chemical that could become drugs.
Nasos: 03:25 How do you see that hopefully changing the world on a grand scale, if your thesis plays out as you hope it will?
Noor : 03:32 The way I like to think about it is like hopefully in five years time when the technology is mature enough to actually produce novel, effective chemical structures, you'd see people discovering diseases and then hopefully within a month you have a good candidate that is ready to go into pre-clinical and clinical stages.
Noor : 03:52 So instead of spending five years just computation, you're virtually coming up with these candidates, you'll shorten that to a month or two. And then you're ready to go to humans.
Noor : 04:04 So the whole cycle could be shortened by at least five years. The effectiveness of the drug could be increased by at least 10X, and hopefully we'll see less diseases and less people die for diseases.
Nasos: 04:16 Tell me a little bit about the story of how you came into entrepreneurship. I understand that you come from an academic background. You got your Ph.D. in machine learning and you also were a professor at a university in Denmark, I believe.
Nasos: 04:28 How did you pivot? Which is quite a dramatic pivot from going from that position in academia to entrepreneurship. What was the story behind that?
Noor : 04:34 I'm not sure it's a big pivot. The big pivot for me was I wasn't really expected to be a CEO. The big pivot was going from a CTO to a CEO. The pivot into the start-up world wasn't really a big jump, because my job at the university started to be more of the administration side.
Noor : 04:57 As a professor, you don't really get to do much of coding or the fun research stuff. You get to do more of administration and fund writing and supervising and these kind of things. Which is more or less what you do in a start-up.
Noor : 05:12 You have to raise funds, you have to manage people, and you have to build your company. You build your research group. But the nice thing about start-up is that you have to move very very fast, otherwise you're basically dead.
Noor : 05:26 That was, I would say, the analogy between the two positions for me. Whether I want to do it in a more slow, relaxed atmosphere where you can't really see the impact of what you're doing directly, because you're basically publishing papers, getting citations, and that's more or less the measure of your productivity. Or whether I wanted to be more or less the same things, but being more productive, more faster, and see a way bigger impact.
Noor : 05:52 And the answer was clear for me. So I moved.
Nasos: 05:56 I want to get into the co-founding story, because as you mentioned in your brief description of what you're doing at GTN, it really is an interdisciplinary project in two pretty diverse fields.
Nasos: 06:05 But how did you get into the field of machine learning in the first place? Why did you end up going down that path?
Noor : 06:11 Oh, that started long time ago. Machine learning specifically started when I was back at my bachelor's studies.
Noor : 06:19 We had five year, I'm originally Syrian so I did a bachelor in Syria. And the program there is five years program where you get to do computer science for three years and then you have two years to choose a major to specialize in.
Noor : 06:34 At my time, I don't know how it is right now, but that was artificial intelligence, software engineer, and I think networking. And artificial intelligence was more so more exciting than the other two. It was naturally like that.
Noor : 06:51 If you want to explore how the world would look like in 10 or 20 years, then you have to go into artificial intelligence. Even if it's not really applicable at that time. So I chose artificial intelligence. And then I did a master in artificial intelligence. And from there it was like natural to go into machine learning and AI and do [inaudible 00:07:09].
Nasos: 07:10 So you were very conscious from that age when you were doing your undergrad that this is going to change the world.
Noor : 07:15 Yeah, exactly.
Nasos: 07:15 AI's going to change the world.
Noor : 07:17 You see the news and you see robotics and computer vision. They were very immature at the time, I would say, but still they were something you could see that will happen in the future. It didn't happen yet, but still. You see the future going that way and you want be in that field. Yeah.
Nasos: 07:36 Talk to me about how this collaboration with Vid came about then. How did you guys end up working together? What were the processes, the conversations, that meant you decided to build something in this space together?
Noor : 07:48 I think it's basically about how EF works. You're put in this room with like 80 other brilliant individuals and from different perspective and different disciplines, and you get to talk to all of them. And see who's the one that you personally like.
Noor : 08:07 It's not just about the idea of background. It's also about how you personally could work together. You keep talking to many people, until you try some of them, and you kind of iterate.
Noor : 08:21 That's basically how we came to work together. But then from the point where we started to work together, we were very conscious about the fact that we are both coming from a very academic background with no industrial experience and no market knowledge. And that we had to work really hard on that before we actually developed anything technically.
Noor : 08:43 We were struggling about whether we should actually make something and go sell it. Or [inaudible 00:08:47] just talk to people and figure out what they want and then try to implement it.
Noor : 08:53 And we were like very aware that the first approach wasn't really the right one. And it comes basically because of our background. So we moved very quickly into just talking to people. We kept talking and talking to people.
Noor : 09:07 We talked to about, I think, 50 or 70, between 50 and 70 different individuals. And just like people working on different verticals and trying to figure out where are their problems and where our technology could fit together and make a big impact, and yeah.
Noor : 09:24 We stumbled across drug discovery. We started digging and it is usually inefficient. It was really a shock just to see the numbers, that the future's expectations and how inefficient the whole process is.
Noor : 09:42 We thought we could do something that is quite disruptive with what we know. Yeah. We started talking to pharma people and they didn't seem to have encountered anything along the lines of what we were thinking. And our solution seems to be quite unique, and even in that very early stage we managed to get some big names in pharma interested.
Noor : 10:05 That was a clear sign for us that there is something really big that we can solve.
Nasos: 10:10 In the initial part of that cost and benefit process, of talking to people, when you obviously had the beginnings of an idea but it wasn't fully formed, I guess, how did you navigate the process of talking to customers and finding out about what they wanted when you didn't have a fully fledged solution to pitch to them?
Nasos: 10:27 What were the processes that you used?
Noor : 10:30 Initially we just reached out for advice. Basically pitch ourselves. I am this person, I have this experience, I'm looking to solve some of your problems, can you tell me about your problems and I'll try to solve them.
Noor : 10:45 That is how it started. But then once we started to narrow down the field that we're working on, we started to have a more focused email, a more focused pitch. Kind of like a few lines that describes in a very high level ... it was high level because we didn't know what it was.
Noor : 11:03 So very high level about the technology and how we think it can solve the problem. And that seems to be effective, somehow. Yeah.
Noor : 11:12 But I think it's important to work on your LinkedIn profile. Because people, before accepting the connection, will look at it and see whether you are credible enough to have a conversation. That was, I think, something important for us as well.
Nasos: 11:25 Yeah, an important little filter just to make sure that they're doing their background checks.
Noor : 11:30 Exactly.
Nasos: 11:31 Then you stack up.
Nasos: 11:32 Tell me a little bit about the process of then going through various cycles and iterations. Because obviously the idea would've evolved many times, I assume, over the course of the months that you were working on it.
Nasos: 11:43 Were there any stages where you really did take dramatic turns? Either before or after the process of actually deciding to focus on that market of pharma.
Noor : 11:52 Like for ideation?
Nasos: 11:53 Yes.
Noor : 11:53 Yeah. Initially we were more focused on application for quantum physics machine learning for compression of neural networks. And we kept, for more than a month, trying to reach out to people within that space and sell them what we were thinking about a product looked like.
Noor : 12:12 But I think EF, and especially OFAP, was pretty useful in that sense, pushing us toward actually is it a big enough market, is it the right market.
Noor : 12:21 So yeah. That's basically why we moved from this market of compression of machine learning ... which I would say it is quite an interesting application. But the market is not as big as we want it to be.
Noor : 12:39 If you do deep enough market research, you will get that sense. So yeah, we decided to move on and keep talking to people from other verticals. And yeah. We moved on.
Nasos: 12:52 In terms of, if we go back in time a few weeks when you were on stage speaking to Reed Hoffman at the event, run by EF, a lot of the questions that you asked were hiring and people and culture. Now I'm going to ask you about those things.
Nasos: 13:08 To what extent is that something that you've done already? And what have you learned about hiring and people?
Noor : 13:12 We keep constantly thinking about those things. I feel like they're the crucial part of every startup. Because it doesn't matter how much money you raise or the idea or the technology, your whatever. If you don't have the right team to execute, you won't be there.
Noor : 13:33 We are now hiring people. We have a couple of ... the ones that we are interested in, they are interested in joining us. But it's also important for us when we interview people to make sure that they have the right mindset. That they are in the startup, in the entrepreneurial mindset, and they're not really expecting very big salaries. They're happy with working with small startup, where the product is not really clear yet, the ideas are not really crystallized yet, and they will be part of the whole process.
Noor : 14:03 Some people get scared of these things. Some people get excited. And you have to pick the right ones. And they have to know. They have to like to anticipate what's going on before they actually jump in to be part of the journey.
Noor : 14:16 We're trying to make sure that everything we're doing, everything we are planning to do, is clear from day one. We're trying to share with them as much as we know about what we're building. And just answer their questions pretty honestly. Basically the main question that they ask that we don't have the answer for, and it is part of their job to help us find the answer, and they should know that.
Nasos: 14:39 Are there any specific questions that you're asking them, to establish whether they are people who would be comfortable in that sort of more chaotic dynamic style of environment?
Noor : 14:49 It's usually like you get an intuition based on how much salary they are asking for and how much they compare what you are offering with what they are getting from big corporate. Or what they expect as a budget for traveling, for the bonuses or these kind of things.
Noor : 15:07 If they're asking for too much, then it's basically you're not in the right mindset.
Noor : 15:13 Also, the working style I would say. So how much [inaudible 00:15:17] is not really working for another very well established company where you go specific times of the day and that's it. Working at a startup is like, at least for the founders, it's been like 24/7.
Noor : 15:34 They should expect that they should work harder than what they usually do for other companies. I guess it's a super exciting journey, and they get to learn a lot. It's going to be different than any other environment that they will have to work with.
Noor : 15:50 But yeah. I think they should be super excited about it. That's all.
Nasos: 15:55 And what about this nebulous word culture that gets thrown around so much and so few people really know what it means. What are some of the concrete things that you are doing, are thinking about doing, within GTN to make sure that the culture is something that you and Vid are setting from the top down. And it's the way that you want it to be, as you scale.
Noor : 16:18 It's actually quite tricky. I think people don't really have a very crisp definition of culture, because there is none. It's basically the personality of not just the founder, whoever joined the team very early on. And then how that reflects everything that's going on within the company.
Noor : 16:38 Whether it's when do you shop, how long do you work at night. Whether it's how much productive you are. What's your style of being there.
Noor : 16:51 And it's very important that you have the right people very early on, because those are the ones that shape the structure. We as co-founders, for me at least, I'm not usually there at the office all the time because I have to be somewhere else.
Noor : 17:06 And these people will be like taking my place most of the time. So it's very important to have the right ones.
Noor : 17:14 I think it's more or less like matching personality, I feel like. Yeah. You might have very brilliant individuals who are willing to join, but personality-wise they might not be the right fit. So you have to make sure that you like working with them. You'll be there all day with them, so you have to be happy.
Nasos: 17:42 Yeah. Sounds like there's a lot of nuance involved in that process of sensing it and making the right people
Noor : 17:42 It's more like intuition, yeah.
Nasos: 17:45 What have you learned about yourself during this process? We've spoken a lot about the lessons that you've learned through GTN as a startup and the progress that you've made on that front.
Nasos: 17:54 But personally, what are some of the things that you've taken away from this entrepreneurial process?
Noor : 17:59 It's tricky. I don't know. I learned a lot, and I really love that. Kind of miss that very steep learning curve and I feel it is super important. Whenever people are sensing that the learning curve starts to be more horizontal than curve, then they should just jump into something else.
Noor : 18:22 I learned that I really love steep curves. I'll say the other thing is that I'm really resilient, so I didn't know that I can be that tough. So yeah.
Noor : 18:38 I don't know.
Nasos: 18:39 Is that in the context of navigating the ups and downs? Is that a specific story that you're thinking back to when you think about that?
Noor : 18:46 Yeah. I'm more like an introvert, so I don't like ... I'm not the kind of person who'll reach out to people or talking to hundreds of people. If somebody says no, then I would go back to them and say why no?
Noor : 19:01 So it was like I wasn't like that. And now I feel like I'm very comfortable doing these things. There's this shift in personality, which I like. It seems like I'm doing it pretty good as well.
Nasos: 19:16 Yeah. Like you said, it was only a couple of weeks ago that you were on stage chatting with Reed Hoffman in front of hundreds of people.
Noor : 19:24 Yeah.
Nasos: 19:25 It's definitely going in the right direction, if you're talking about that.
Nasos: 19:28 What advice would you give to female founders? I know it's something we think a lot about at EF, getting more female founders on board, etc. What advice, if any, would you give to them?
Noor : 19:40 Well, to be honest, I don't really like to think about ... I don't like the discrimination between males and females. When you want to take a decision, just take it regardless.
Noor : 19:51 Because there's fundamentally no distinctions. If you want to follow your dreams, if you want to make an impact, if you want to do something, don't think about yourself as a female. Just think about yourself as a human being who wants to do something. And do it.
Nasos: 20:08 Just simply find a frame.
Noor : 20:09 Yeah, exactly. We're all basically humans. It's like ... yeah.
Noor : 20:14 When you talk about startup, doing things or having a dream, pursuing your dreams, there's no differentiation.
Nasos: 20:23 When you think back to before you were applying to EF or when you were thinking about applying to EF, was there anything that was stopping you? And what helped you break through it?
Noor : 20:37 Something ... it was basically logistics. It wasn't really something that's stopping me in terms of wanting to do it. It's more like I have my daughter and my husband, in Copenhagen, and moving to London, leaving my daughter behind. It was a little bit of ... I wouldn't say initial, but it's something to think about.
Noor : 20:58 And how easy, what was it like to move her to nursery and how quick we can do this thing. It was more like practical things rather than fundamental things.
Noor : 21:08 And it turned out to be fine. Yeah. You just have to take the first steps and the other steps will follow and things will move on. Yeah. It will resolve, everything will be resolved and it's fine.
Nasos: 21:21 So your advice to anyone who is thinking about starting a company, it's a common sort of thing that people will sometimes come up with. It's not the right time, for whatever reason it might be. Family purposes, etc. You obviously had a unique situation.
Noor : 21:36 It never is. Yeah. And that's applied to everything. Whether you want to get married, whether you want to have children, whether you want to start a startup. It's always not the right time.
Noor : 21:45 And it will never be. So just ... yeah. Make up your mind and
take the decision. Go on with it. Live with it.
Nasos: 21:52 I think that's a pretty great note to end on. Noor, thanks so much for coming on the show. It was really great chatting to you.
Noor : 21:52 Thank you very much for having me. Thank you.