#12 David Hunter of Optimal Labs on Revolutionising Human Nutrition
By Nasos Papadopoulos, EF Head of Content
David Hunter is the CEO and Co-Founder of Optimal Labs a company that applies cutting-edge deep reinforcement learning to create intelligent autopilots for farms, improving the efficiency, reliability and quality of food production.
High tech greenhouses can produce 10-40x yield of traditional farming and they’re universally scalable and deployable anywhere. And the implications of AI controlled farms are potentially huge – with higher quality food, produced faster and more reliably, people will be able to live better for longer.
David is one of the famous cases of someone who did EF twice – he was on EF4 and after met his Co-Founder Joao on EF7 the idea for Optimal Labs became a reality and before they knew it, they were following farmers around greenhouses in Holland and building solutions to their problems.
He studied Aeronautical Engineering at Imperial and trained as a pilot in the Royal Airforce before going into the banking industry, where he eventually ended up running Deutsche Bank’s quantitative strategies team where his algorithms traded >5% of the European stock market volume.
After going through EF4 he took some time off to earn a Masters in Deep Reinforcement Learning at Oxford before returning to join EF7.
In this episode David and I discuss:
- Why he left his well-paid job at Deutsche Bank to start a company
- What he learned from doing EF and how he applied this knowledge second time around
- The challenges he’s overcome along the way and his tips on hiring and seeking advice
This was a fascinating conversation that will not only give you an insight into the future of nutrition but will also show you the resilience and continuous learning that entrepreneurs have to apply in their lives.
Why I Went to the Edge of Science Before Building My Company
What I Did Differently Second Time Around at EF
Why I Joined EF
03:23 Dave, welcome to the show!
David: 03:24 Hi, how's it going?
Nasos:03:25 Not bad at all. Pleasure to have you on. It was real privilege to have a quick chat with you before and get a little bit deeper into what you're doing at Optimal Labs. To kick us off with, just give me a quick 30 second intro into what you're doing and how the world is going to change as a result.
David: 03:42 Yeah, I guess the mission of the company is to increase the availability of safer, healthier food by deploying AI controlled greenhouses outside of every city on Earth. That's our fairly unambitious mission. Starting around 60 years ago there was this big green revolution and ever since then this combination of new lab made fertilizers and genetic engineering that has been increasing food yields and has been saving us from starvation is a rapidly growing population that we've had but at the expense of food quality and also severely constraining our resources such as water and this kind of thin. So, really what we want to do is enable the population to grow from here but also reversing this downward trend on food quality and we see that growing in controlled indoor environments is one of the big ways to do that.
Nasos: 04:44 And so how will what you're developing change the way that the world consumes food in a practical sense? I mean, we spoke before about the implications on yields and consumption patters, etc., etc., in a very tangible way, what would it look like for someone living in an urban environment and how would their life be different? How would the quality of the produce they're eating be different?
David: 05:08 Yeah, well, in terms of food quality, you're right in talking about an urban environment. We're getting an increased amount of urbanization and that's likely to continue. The majority of people moving away from rural areas into urban areas and the majority of urban locations around the world are poor for growing crops, generally. Either they're in very hot climates or cold climates and certainly with increased climate change and this kind of thing, climate conditions become more variable and cities tend not to be in locations that are good for growing and so what this means really is that crops tend to be grown a long distance away from where they're consumed and what that means is that crops are really grown for transportability right now.
David: 05:59 So, you're growing a tomato in a field in the South of Spain or Mexico or something like this, it often has to be transported literally thousands of miles before it's consumed, especially in places like North America. And so what this means is you end up with basically a produce which, where the nutritional quality has been sacrificed. On top of this a lot of pesticides tend to be used to grow that produce that's only available certain times of the year and this kind of thing. So, if you can imagine now, we fast forward and we have huge, large scale greenhouses and or farms outside cities. You can pretty much control every aspect of the growing climate.
David: 06:45 So, you can really grow food and optimize it for nutritional quality. Theoretically it can be harvested and consumed within hours of it being harvested and so you don't suffer this nutritional degradation because you're growing in an enclosed environment you can grow with minimal to zero pesticides and there's a whole host of other benefits as well like reduced water consumption and this kind of thing but as consumer it means that you'll have access to nutritionally optimized food grown all year round and grown hyper local to where you're consuming it.
Nasos: 07:20 Let's just take a step back and go back to the beginning of your career and how you were thinking about what you wanted to do. You were working in banking, you were running towards the end of that stint, a quantitative team at Deutsche Bank, if I'm not mistaken. How were you thinking about your career back then and what was nudging you towards doing something different?
David: 07:47 Yeah, it's funny because when I was a kid I used to get told by people, "Oh, there's no such thing as a straight forward, linear career and everyone has switches and this kind of thing, u-turns." And I dismissed all that and now when I compare what I've been doing, compared to some of my peers, I've realized I've probably had the most variable career out of all of them. When I was younger I wanted to be a pilot so I sponsored through university by the Royal Air Force and I did part of my flying training while I was at the university and that was my childhood dream and then towards the end of my degree I realized what it meant to actually be in the Air Force and it wasn't as glamorous as Top Gun which is what I was hoping it would be like.
David: 08:38 So, in the end on the balance side of it, it wouldn't be right for me so I made this last minute decision to go into banking and I was very fortunate because I was put onto the algorithmic trading desk and allorhythmic trading was just really starting as a thing in European equities trading and I think I was probably the first graduate to be hired onto an algorithmic trading desk. And I just grew with that. So, I have this very fortunate right place, right time career within banking. Since I was a kid I've been programming computers and I think I was the only trader on the trading floor who knew how computers really worked and so I was able to bridge a gap between technology and human decisions within trading and so I ended up, at a very young age, running the algorithmic trading desk at Deutsche Bank.
David: 09:35 I did that for a few years but it wasn't a particularly hard decision. So, after doing that for seven years, I think, most of the trading had become automated. So, when I started in training, I think something like 30% of all trades were executed by machine and 17% were executed just manually by people. By the time I left Deutsche Bank we had managed automated, pretty much, all of the trading volumes. So, something like 99.5% of all the trades within Deutsche Bank were executed by our systems that my team had developed so we were only up, our servers were trading something around 5% of the European stock market every day and it just shows you the power of machines.
David: 10:28 A small team of 10 people could control all these trade of 5% of the stock market each day but at that point there's not much room to grow from 99.5% to 100%. So I knew I had to leave and so I left and I went through this very up and down journey of entrepreneurship where I was failing a lot, working on the wrong ideas, working on the wrong ideas with some really good people and eventually I ended up with Oxford doing what I'm doing today which seems to be going well.
Nasos: 11:05 So talk to me about the decision to join EF in the first place? Because you're one of the rare candidates or cowart members who's actually done EF twice. So, what was the factor that pushed you towards applying for EF the first time around and how do you remember that? The initial experience on your first cowart which was the EF-4.
David: 11:29 Yeah, I remember looking at [why combinator 00:11:33] and part of the application procedure was you had to say who you're working with as a co-founder and you had to tell them about the relationship that you and your co-founder have had and how long you have been working on things together and there was always something a bit irritating about that because in the UK we don't have this or we didn't have this natural start up ecosystem where you could meet co-founders and I think about this a lot, back then, Google, Facebook, these kinds of companies weren't that big in Europe. So, in the US you could go and work for Google, you'd be surrounded by all these really smart people who are interested in technology and on the side you could work on side projects.
David: 12:19 It was like a natural, organic way to think of start up ideas and then leave and do your own thing but in the UK most of the people I know who did engineering and other technical subjects, their goal was to go and join finance. It wasn't to start companies. It wasn't really a thing. Doing start ups in the UK wasn't a thing and so we didn't have this natural way for people to meet and I was thinking about it a quite lot and I got introduced to Matt Clifford through a friend of mine and he had also been thinking about the same thing. So, he told me about EF and it just seemed like a great fit. So I joined not really knowing what I was going to be working on, other than this seems like a really good place to meet other people who are also looking for co-founders and looking to start companies. So, that was the reason why I joined the first year of cowart.
Nasos: 13:16 When you look back on that experience, obviously it turned out that the company didn't come out. That release of not company that you continued. What did you most learn from the experience and didn't you also consciously try to apply some of those letters when you came back second time around on EF-7?
David: 13:33 Yeah, big start. When I joined EF-4 I knew I wanted to start a technology company but now looking back on it and saying you want to be a technology entrepreneur is a bit like saying you want to be an author. It's not enough. It's not specific enough. What is technology entrepreneurship? It's all about basically creating enough tools that you're going to sell to satisfy some unmet demand or need or something. You think about the definition as extremely, extremely broad and if you think about technology companies, they're also extremely. On the one hand you have operational intensive companies like Deliveroo. On the other hand you have deep biotechnology companies, AI companies and this kind of thing and you can see that these, relatively both technology start ups, these are very different types of startups and so when I joined EF-4 I hadn't really given much thought but I think my thinking was still quite basic and I basically ended up working on, in hindsight, was a wrong idea for me but with an amazing co-founder.
David: 14:48 So, we can talk about that in a second but yeah, I ended up working on the wrong idea with basically the right person and it didn't work out and it was failure and there was some lessons to be learned and this kind of thing and when I came as an EF this time around I was very clear about what kind of company I wanted to start. Although I didn't come into EF-7 with Optimal as an idea I came in with a very clear idea of the kind of company I wanted to start and that meant that I was able to find the right co-founder and do things in the right way from the start and not get to distracted by these seemingly great ideas that weren't quite a personal fit with me. So I think this founder idea fit is really important and it helps if you've thought about that before you come into program.
Nasos:15:50 Sure. So, how did you conceptualize that type of business that you wanted to start when you came in the second time around? What was the line? How were you thinking about it in your head specifically?
David: 15:59 Yeah, I went through this very convoluted thought process and what came out was very simple. Which was basically, I wanted to apply new science. So stuff that's really just being discovered within science and apply that to build some kind of useful technology for people to use and so really I wanted to focus this, it's a bit of an old school definition but I would call it technology transfer. So, really, let's get to the edge of some field of science and then once we're at the edge let's see what kind of possibilities there are to actually build real world products leveraging that technology and I realized that this, for various reasons, this was a perfect role for me. And I can just see myself doing this as vacation. I saw it as a vacation here. This is a useful role for someone to fill in society. We need lots of people to do science and advance human knowledge and we need lots of people to take that bit of advancement and turn it into useful technology.
David: 17:08 And so once I narrowed that down it became a question of, what technology? What are of science do I want to get to the edge of? I'd been interested in machine learning for awhile because when I was doing algorithmic training we were applying a lot of machine learning tools to help us predict price and volumes and other characteristics of the stock market and so when I left EF-4, after doing that bit of thinking, I decided the best next step rather than me start a company right away is let's get to the edge of this field. So, I went back to university, I did a masters in computer science and I spent most of the year doing research with the reinforcement learning and the gamble paid off. Once you get to the edge you do see possibilities and that's why I joined EF again and started that off.
Nasos: 18:01 How deliberate was this process then that you embarked on after EF-4? Reflecting on what you learned and then going into this specific field and then looking in the specific area after that. How strategic was it?
David: 18:15 It was extremely deliberate and strategic. So, after reflecting on EF-4 I realized that the idea I was working on was the wrong idea and what I really wanted to work on was stuff, this technology transfer thing were you take stuff that's on the edge of science and new discoveries in science and then see how you can use that new knowledge to build useful technology and really build a company, build products, that are on the edge of technical possibility and where the risk is in technology. There's little risk in the market or having good operations for example. Which I've said the other two types of business risk and so once I made that decision it was a deliberate process of looking into the different massive areas of science, which could potentially have high engineering leverage because there's a lot interest in things happening in science but only a small sub set of those have real world engineering leverage.
David: 19:21 And so, once I'd done that, in hindsight now it seems extremely obvious, but when it's a machine learning I did a masters so I could get to the edge of that field. Initially I planned on doing a PhD as well but I realized that as founder of a company you don't need to necessarily have PhD level knowledge. You need to be thinking about how you can apply this, you can build new technology and solve a market and this kind of thing. So, you need to have a bit more general skill set and so I decided to stop after the masters and then stopped and all but it was a very deliberate process.
Nasos: 20:04 And so you mentioned first time around that you were working with a great person on the wrong idea. What were you looking for in your co-founder second time around? And when and how did you realize Joel was the guy to go ahead with you on this?
David: 20:16 Yeah, that's a great question. This leads me into this founder idea fit kind of thing. So, when I joined EF-7 there were a lot of people. Generally the way a founding team comes together is, there is one person who is considered to be a bit more technical and one person who is considered to be a bit more operational and so that operational person naturally falls into the CEO role and that more technical person falls into the more CTO role. So, when I joined EF-7 I was looking to have that operational role be it in a very technical company. Where even the CEO needs to have a high technical ability and when I was looking for someone even more technical than me, someone who had done a PhD or something like that or had extremely deep knowledge in reinforcement learning and machine learning.
David: 21:10 So, when I joined EF-7 that's what I had in mind but when I joined there were people who were trying to ask me to join them as the CTO. So, it just shows you I could have been the CTO of a less technical company with more operational risk by partnering with an [xmachenzie 00:21:32] or BSG associate or something like this or I could have been a CO at a very technical company but I knew I wanted to be the latter. So, that made my decision relatively easy because my co-founder, Joel, he was doing a PhD in robotics EPFL and he had that real deep knowledge and was interested in nutrition and everything. So that was a very obvious choice.
Nasos: 21:57 We were chatting about, before we started recording, all the cool projects that you've done over the last few months and the awesome progress you're making but every entrepreneur will know that there are ups and downs to the journey. What's been the most difficult challenges you've had to face and how did you pick yourself up after it?
David: 22:16 Yeah, well I can honestly say I think, personally speaking, the most difficult thing was after EF-4 seeing how I had this great job in banking and this fast career path and then I left and joined EF-4 and thought, "I was greater out with my training so surely I can start a company." And reality gave me a kick in the face, sort of thing and I went through a good couple of years of failure basically before eventually going to do the masters and doing research in machine learning and then joining EF-7 and starting Optimal. So, and it's easy to gloss over that. That was actually quite a low point, personally but I think once you've made a decision from first principles about what you want to do. It's very hard to be shaken from that. So, I think once you've convinced yourself, "This is something I want to do. This is a role that I want to play." You just keep going after it but it wasn't clear that it was going to be successful at the beginning and with the masters again, I didn't know whether I was going to find any decent ideas off the back of it but you do.
David: 23:38 And what's actually quite an interesting little side note is that I said, in EF-4, I was working on the wrong idea with the right person. Well, the person I worked with on EF-4 as a co-founder we became great friends since then and actually last week he accepted an offer to come and join us at Optimal. So, it's awesome I get to work with him again and I read something once that, "you should only work with people who you would work for," and that's true of him. So I'm just delighted that he's going to be a part of the journey. So these things have a way of working themselves out if you're persistent enough.
Nasos: 24:18 How are you now thinking about hiring processes as well? 'Cause you mentioned one of the challenges you're currently facing is trying to get more people on board. Is that a that you're using of, "would you work with this person" and bring them on board? How are you approaching the process? What kinds of things are in line now?
David: 24:33 Yeah, we have this very strong view that the best people for a job are not necessarily the people with these full academic credentials and so, we're hiring reinforcement learning engineers, machine learning engineers, software engineers and we have a very ... our process is very skills based and we're valeted by their skills and cultural fit and we're testing them, so to speak, as part of the interview process. I guess, we're more merit based and we're trying not to look too much at which universities they went to and what marks they got academically because in a start up environment ... in academia someone gives you the work. Someone says, "Here's a problem. Go and solve it."
David: 25:28 In a start up environment you have to figure out the problem and then you have to solve it and figuring out the right problem to solve is actually harder than solving it often. So, making sure you're working on the right thing and so often just because you're good in academia doesn't mean you're necessarily good in start ups and so we're trying to take this merit based approach and the value of people based upon what they can do rather than what credentials they have on their CB's.
Nasos: 25:56 Sure. I'd love to spend more time going into more detail as to the awesome stuff that you're doing but in the interest of time, I know you've got to run soon and let's finish up with a little bit of the reflection on the experience for you personally. So, what have you most learned about yourself over the course of this journey from leaving Deutsche and starting a company in EF-4 and then having to make adjustments, starting off to now obviously the great progress that you're making. What have you most learned, personally, through the experience?
David: 26:27 I think starting a start up is a lot about the search for truth. You're trying to build a product and the market will decide if it's any good or not. So, it doesn't matter what your assumptions are. What matters is what your customers say and this search for truth is important and that goes for the market but as well as to yourself. So I think it's important to not have too many assumptions about what you're going to be good at. I've got this quite technical background, I've got a computer science degree and this kind of thing, but actually at Optimal I'm doing more of the sales role and operational role and this kind of thing but I've actually found that I'm good at it and I enjoy it.
David: 27:08 It's something that I didn't think I would be doing going into it but I'm just embracing it and running with it and I'm bringing in other people and my co-founders are doing the technical stuff. So I think it's being relatively open minded and even with yourself, trying not to fight yourself too much. If you find that your naturally drawn to some particular area of a business then you should just go with that and you should hire in great people to do the other things and trust them to do that job. So I'd say, yeah, that' the biggest thing I've learned. I'm becoming more of this operational person rather than a technical person I thought I'd be and that's great.
Nasos: 27:59 Great. Well I think the aspect of developing oneself is something that so many entrepreneurs say is critical to success. I think it's a great note to end on. David thanks so much for coming in. It was a pleasure chatting to you.
David: 28:08 Thanks so much.