SDS 577: Scaling A.I. Startups Globally

Jon Krohn: Welcome to the SuperDataScience podcast. The most listened-to podcast in the data science industry. Each week, we bring you inspiring people and ideas to help you build a successful career in data science. I’m your host, Jon Krohn, thanks for joining me today. And now let’s make the complex simple. 

Jon Krohn: Welcome back to the SuperDataScience podcast. Today’s guest is the sensationally, successful AI entrepreneur, Husayn Kassai. Husayn, co-founded the machine learning company Onfido in 2010 while he was an undergraduate student at Oxford and served as Onfido’s CEO for 10 years, overseeing $200 million in venture capital raised, the team growing to over 400 employees and the client base growing to over 1500 firms. A little bit on, Onfido itself, it’s a company that enables businesses to digitally onboard users, such as Uber with its drivers or couchsurfing.com with its hosts and its guests. It allows them to onboard these kinds of users more rapidly and safely by using machine learning models that both verify government-issued IDs are legitimate and then compare those IDs to users faces. In addition to founding that company Onfido, Husayn completed his BA in economics and management from the University of Oxford and served as the full-time President of the Oxford Entrepreneur’s Student Society, which is how I got to know him in the first place. 

Jon Krohn: Today’s episode is non-technical and will appeal to anyone who’s interested in hearing tips and tricks for building a billion-dollar AI startup from scratch. In the episode, Husayn’s saying details, tips for deciding on whether you need co-founders, how to choose your co-founders if you need them, how to find product-market fit, how to scale up a company, how to identify startup opportunities, why there’s never been a better time than now to found an AI startup and a behind the curtain look at his next startup, which is currently in stealth. All right, you ready for this inspiring episode? Let’s go. 

Jon Krohn: Husayn yes, I get to catch up with you my old friend on air, on the SuperDataScience podcast, this is so exciting. I’ve been looking forward to this day for a long time. How are you doing Husayn? Where in the world are you calling in from? 

Husayn Kassai: I am very well. I’m in Kentish Town, North London, and really, really pleased to be on. 

Jon Krohn: Nice. Yeah, the audience is in for a treat today, for sure, because yeah, you’re a tremendous person. You exude this calmness that I come away from every experience with you, feeling like everything in life is great and under control. And so I hope that the audience comes away feeling the same way from this episode and inspired by the things that they can do by scaling up AI applications, even a fraction of the scale that you have. We’ve known each other for a very long time. I was at Oxford studying from 2007 to 2012, and I started getting involved with the entrepreneurship organization of the university called Oxford Entrepreneurs in 2008. And I served in leadership roles there up until I left the university in 2012. And I did things like, I was a VP of the organization. And so doing things like interviewing the president. 

Jon Krohn: So the president was a pretty cool job that people could have at OE that we’d actually pay you a salary. Typically people would take a year off either take a year, interrupt their studies for a year or take a year between their studies and say, launching a business or getting involved in some other commercial opportunity. And so in 2008, you started getting involved with Oxford Entrepreneurs as well on the committee. And in 2011, I interviewed you for the president job and you were a shoe-in, you were president for that year until 2012. And I remember not only were you an amazing leader of the organization, but there was a specific day we were in the business school in Oxford and in the front hall and you were telling me about this idea you had. You’d been applying for different job opportunities and you couldn’t believe how much duplication of effort there was. For every job that you applied to you had to create a new profile. You had to answer new questions. You had to provide the same verification documents over and over again. And so you said to me, “Jon, I think that there’s an opportunity here.” And then, many years and $200 million in capital raised from the likes of TPG, Salesforce and Microsoft, you created, you were the CEO and co-founder of this massive tech startup Onfido that started from that idea. 

Husayn Kassai: Former CEO, correct. 

Jon Krohn: Yeah, yes, yes you were. And yeah, it’s so amazing to have you on, and you can tell us about that journey from that kernel of an idea to, I remember things like you telling me how the big banks, investment banks that were the kind of job that you thought might appeal to you. Everybody’s working in spreadsheets and you thought there’s got to be a better way to be working. There’s got to be ways to be more efficient. There’s got to be opportunities to automate and innovate in the world and you wanted to be part of that. And, yeah, I don’t know if you want to just tell us about the beginning of this journey where you were a student, then you became the president of this entrepreneurship organization and you had this idea, this opportunity, and you saw the potential in software and automation and yeah, so what happened? How did you have the courage to take those first steps? 

Husayn Kassai: So during the time at Oxford Entrepreneurs, I got to, along with some friends, we got to see how the world is changing quite rapidly. And there are an increasing number of tools available that make it a lot easier to start a startup and try out new things. And so we got to see that firsthand. And as you can imagine, there’s a bit of excitement that comes with this ability for you and a few friends with some laptops and internet access to build something that has a potential. Very, very small likelihood, but has the potential. It is possible for it to potentially touch everyone in the world or have a chance to scale globally. With all that excitement, then afterwards going and working at essentially investment banks, me and my two co-founders, we sort of, it was a bit of a [inaudible] in that, sitting down spreadsheets or typically not using the latest software even, it was Windows 2000 or whatever it may have been. 

Husayn Kassai: And so the first half with the Oxford Entrepreneurs introduced us to technology and what can happen and how the world is completely changing. And then the second half being part of these larger corporate environments, it was a bit of an understanding around the culture, the norms and habits that guide constructive innovation and teamwork and things of that nature. And we kind of came to the decision that we ought to do a start of ourselves. I did a one week internship, one of my co-founders did the whole summer internship and our third co-founder, he actually spent two years at an investment bank before he joined full-time. It was wonderful. And we learned a lot, but we knew that we wanted to set up a culture where people would very much enjoy coming to work and not lose that like passion and excitement that they don’t need to spend two to three years working on Excel spreadsheets before they’re given a chance to do something. They can join and from pretty early on, they have autonomy and sort of a path to really start making a contribution. 

Jon Krohn: Cool. And then, so the specific idea behind Onfido that the three of you co-founded you, Eamon and Ruhul, I know from something I read online, actually not something that I knew about you personally from before is that, your family had some identity verification challenges when they moved back to England from Iran. How did that identity verification challenge become part of the genesis behind the idea of Onfido? And why is this identity verification challenge important? 

Husayn Kassai: Specifically, when I turned 10, my parents moved from Iran, they came to the UK and I just remember, even though I was quite young, they struggled to open a bank account and rent in their own name. Growing up, at a young age, I realized how important identity verification is essentially being able to prove that you are who you claim to be. And that it’s all underpinned for the most part, or it used to be before Onfido and our approach. They used to be underpinned by credit bureaus. And then when I looked into the credit bureaus, I kind of saw that this is essentially, these are centralized databases of everyone’s date of birth, name and address. And that whole approach is about 150 years old. And the two main issues are, part of it is that it excludes half the world adult population who are under banked or unbanked, or maybe have recently moved to a country and therefore don’t have a credit footprint such as in my situation. 

Husayn Kassai: Or the other half of the world’s population, such as us, where we do have a credit footprint. It is so easy for fraud source to commit identity fraud and pretends to be you. So all that fast and all those sort of unnecessary hoops to jump through for this illusion of security. And yet you’re not even getting security, because it’s so easy for force to institute it. And that was clear for me and what a fundamental problem we have here. And in parallel, I got quite fortunate in working with our technical co-founder in particular, to see how machine learning offers a path forward for there to be a superior way to verify people or who they came to be remotely. 

Jon Krohn: Super cool. And so part of what allowed Onfido to disintermediate, to provide a technological solution that this 150 year old credit footprint industry couldn’t, was AI, was data science. You were able to use machine vision to try to predict whether a given government ID is real or is fake and then subsequently you can match the person’s face to their ID. And so as you say, this goes beyond just looking at somebody’s address, which could easily be falsified. How did that come about? How did you realize that there was this potential AI driven solution? 

Husayn Kassai: We had our technical co-founder. It just so happens that is university thesis was using machine learning to spot wildlife in 25,000 photos of the jungle. And essentially, he was explaining it to me and he explained how these models can be trained to recognize patterns. And then when you feed it more data, it is at first human assisted machine learning and then can do it at scale more effectively and so on. Then that’s when I started to say, if there were to be a government ID, can you see these patterns? Because this is structured data. And he said, “Yeah, this is like a very obvious thing.” I said, “If it’s so obvious, no one doing it,” that it really needs… If there were to be a startup that solves this problem, it can solve a global problem that pretty much every human on the planet faces. And that is in ability to verify that they’re who they claim to be online in a digital way. And then I went into and I continuously have done, I’ve got a habit of going on about how much of a broken system we have as far as identities is broken and getting worse. And how the credit bureaus are fundamentally predicated and historically have been around exclusion of people. 

Husayn Kassai: And whereas this new age requires us to rethink it and develop a way to make it easy for more people to prove that they are who they claim to be and at a much lower cost and a much greater security. And that whole process and sort of going to a union in 2008, it’s just off the back of the financial crash. And so throughout that whole period, all these remodels or all these models were being rethought as well. And that kind of was the early stages of thinking about new FinTech and new ways of doing finance and a big barrier to a new approach to finance beyond sort of traditional incumbents, mainstream banks and others has always been this identity and compliance components. The need was always there. It was always broken at least in our minds, but it was becoming even more painful. And yet it was great to see how advances in machine learning were able to, it’s not necessarily advances in machine learning, it’s just the affordability of servers and being able to run models. And all this to my earlier points around the tools to actually get a system going became so much more accessible and affordable that we could start playing with prototypes. 

Jon Krohn: Eliminating unnecessary distractions is one of the central principles of my lifestyle. As such, I only subscribe to a handful of email newsletters, those that provide a massive signal to noise ratio. One of the very few that meet my strict criterion is The Data Science Insider. If you weren’t aware of it already, The Data Science Insider is a 100% free newsletter that the Super Data Science team creates and sends out every Friday. We pour over all of the news and identify the most important breakthroughs in the fields of data science, machine learning, and artificial intelligence. The top five, simply five news items. The top five items are hand-picked, the items that we’re confident will be most relevant to your personal and professional growth. 

Jon Krohn: Each of the five articles is summarized into a standardized, easy-to-read format, and then packed gently into a single email. This means that you don’t have to go and read the whole article, you can read our summary, and be up to speed on the latest and greatest data innovations in no time at all. That said, if any items do particularly tickle your fancy, then you can click through and read the full article. This is what I do. I skim The Data Science Insider Newsletter every week, those items that are relevant to me, I read the summary in full. And if that signals to me that I should be digging into the full, original piece, for example to pour over figures, equations, code, or experimental methodology, I click through and dig deep. So, if you’d like to get the best signal to noise ratio out there in data science machine learning and AI news, subscribe to The Data Science Insider, which is completely free, no strings attached, at Superdatascience.com/dsi. That’s Superdatascience.com/D-S-I. And now, let’s return to our amazing episode. 

Jon Krohn: Yeah, it was around the time that you guys founded Onfido. It was in 2012 that we had the first deep learning model that became a worldwide sensation. We had this AlexNet architecture that was this machine vision algorithm that could recognize images and classify them correctly to a much higher degree of accuracy than any other algorithm on the planet. And as you say, the big thing that allowed that to be possible was that compute had become cheap enough, storage had become cheap enough and the tooling had become available to start to allow us to relatively, easily make use of the cloud compute that we need to be able to train an algorithm like yours to distinguish whether a government idea is real or fake. And yeah, super cool. Amazing that it happened so organically. I mean, how fortuitous to have at that time where, because of the financial crisis, people are looking for new kinds of solutions to old problems. Machine learning models have just become accessible enough and powerful enough to be doing the kinds of things that you and your co-founders had dreamt up. And that you had that expertise that you just, you have this innate pragmatism and commercial acumen that makes you an outstanding CEO. And then you also had someone like Ruhul who had this machine vision background. And so together yeah, you were able to get Onfido off the ground. So what was that like in the beginning? What was the story around developing the tech versus maybe looking for seed funding from VCs? How did the kind of the initial seed stages play out for you? 

Husayn Kassai: And so, yeah, there was my third co-founder Eamon who also came specifically on the operations and the scaling side, he helped a lot. The initial phases were pretty much as soon as I graduated, dropping out was on option and I never even thought about it. I’ve got Middle Eastern dad, it just doesn’t come into the vocabulary. Graduated, started working on the prototype, my technical co-founder Ruhul at the time he had his job and he had the mortgage. I had an agreement with him. We get to our first client, you work evenings and weekends just help me on the prototype, let me sign a few clients. When we get funding, when we get paying customers, I need you to commit that you’re going to resign and jump on a full time at a much reduced salary, but this is at least these are the proof points you’ll need. And we kind of achieved that within a year. And that was perfect because when that happened, after that first year, our third co-founder, he just finished his exams. And so he was able to join full-time. It was like about a one year period where the other two part-time had to sort of help evenings and weekends. And there was me full time. And then after that first year, the two of them came and then there were three of us. There were challenges, you look 18, 19, I was told that I look younger than my age at the time. 

Jon Krohn: You did. 

Husayn Kassai: Trying to go and sell a security product to banks and others, when you look 15 and yet you’re talking about machine learning, which is, it wasn’t in the disclosure vocabulary at the time. And you essentially go to someone at the bank and you tell them, “By the way when people come in and sign up an account, they can do it from home with the smartphone, and they’d show their ID and to show their face,” it was too wacky for them. It wouldn’t have worked. So we had to go find people who were of a similar mindset. And that ended up being exactly startups like us. A lot of, again, good timing. We were lucky that it was a rise of the trust marketplaces. So if you want to nanny, choose your doctor, Tina, coming to your house that whole on demand wave, then there was a- 

Jon Krohn: An Uber driver. 

Husayn Kassai: Precisely. Then you had the sharing economy. It was called at the time trust marketplace, where you’re sharing a car with someone like blah, blah, car or sharing a home with someone like that, couch surfing. Then we started to see in 2013, 14, the rise of like the FinTechs. We were lucky in that right timing for the right customer cohorts at the right stage. And then when we started to sell into FinTech and financial services, at that point, the scale and the volumes became so high, that our models were consistently able to prove scientifically that we are able to catch fakes better than the human I can. And so at that point it started to become… It was at early stages, but eventually became exponential. And that sort of really matched very well, the growth of the FinTechs. And in many ways, it’s, would’ve been hard to envisage trust marketplace or the FinTech ecosystem being as big as it is were it not for tools such as Onfido and similar ones. 

Jon Krohn: For sure. Super cool. And then, so that kind of gets us through the seed stage and finding those initial customers, then what was scaling an AI company like this like? What were the easy parts? What were the hard parts? You went from being based in London and having your team entirely there to then making the decision to have a large part of the company in the Bay Area. And so yeah, these kinds of decisions. Yeah. Like I said, what were the easy parts? What were the hard parts? How did you scale this AI company from your first few customers to raising $200 million? 

Husayn Kassai: This is just about 2015, where we had a basic product. It’s only fed say’s, called call it basic, but comparatively there weren’t… We were still strong and better than alternatives, but compared to where we’re at today, it’s first is quite of a basic version. We straight away recognized that the US market is the one that you need to win. And we actually invited some of the key investors. At the time there were mostly ancient investors to say, “Look we want to do this in the US. And we need to, because this is essentially a race for data. And we need to be able to cover more and more IDs globally to build the best models, because it’s not good enough just doing French IDs and new British IDs. Well, you need to be able to do the Maltese, or Indonesia or the Vietnamese and so on and so forth.” We said, “We want the ability to sign clients that are global.” And most of these global companies are based in the US and they don’t want to buy from a UK company. They want to buy from a US company. And there was pushback. We were kind of told, “Look, the ballpark number is you have to get to 20 million in revenues before you should consider the US.” And at the time our revenues was close. This is 2015, it was closer to like 20,000 pounds, there was a little bit of a gap. 

Husayn Kassai: But as it happens, most of my family are based in California. And so we kind of launched in the US without telling the universities at the time. But it was a slow start. 2015 was the first year we tried to launch the US, had it the first team of about three people and then 2016, 2017, 2018, just a very, very difficult path to coming out as a UK startup, trying to go through the same issue of selling a security based product to US institutions. It took a good three, four years and it was 2018 when we actually broke into the US market. And we started to become respected as a scale up and started to win some of these very large brands, as well as ultimately the banks. But it took a long time. It took much longer cracking the US than it did even setting up the UK. But it was never an option, like we made it clear to everyone, basically if we were told, win the US and lose the rest of the world or vice versa without blinking twice, we’d say win the US. Not winning the US was never really an option. 

Jon Krohn: Cool. Yeah. Incredible journey. Great to hear from you, kind of from where you and I left off. And I guess actually that was the last time I think we’ve seen each other in person was 2012. Now 10 years ago, we’ve caught up a couple times, otherwise since, but in person, I think it’s been 10 years. We’ll have to fix that soon, but amazing to hear from where I left you off, your journey to growing this massive company, but you’re not the CEO anymore. You have now gone off, maybe you enjoyed some gardening or something for a while. I don’t know how much you could do that you probably can’t sit still. 

Husayn Kassai: Conceptually, it sounds good. I have not yet done it. No, but when it came to pretty much the 10 year mark, we brought in a new CEO and the plan is essentially to run an IPO and run towards an IPO. And I took a break and then in the summer started a new company. We’re probably eight months into the new venture. But previously you’d also asked me about the kind of this scale up journey in your last question of what that was like. It was for me, fundamentally value creation comes down to the team. And the what set us apart from everyone else, because it was relatively competitive at the time was fundamentally the team. And so our sort of other co-founder Eamon, he did a lot of heavy lifting on the operations side, but together we focus a lot on the team and read about what makes a great culture. 

Husayn Kassai: We were involved in interviewing every single person who joined and constantly recognize that as the most important thing, essentially who you are hiring and having a really, really good process for that starting with what problem is it that you’re looking to solve and what kind of profile do you need and then drawing in advisors or experts or the best is usually other founders who’ve gone through it to help with that. And once they’re on board, how can you develop them? How can you help them do their life’s best work and enjoy it more than any other career or job they’ve done in the past and contribute more than they ever have? And how do you do that? 

Husayn Kassai: And ensuring that the culture is strong, the norms and habits and what makes it ultimately for people to want the company to achieve its mission, beyond every single person’s a shareholder or share option holder and then ultimately a shareholder. What are all the things that are possibly doable that we can do to make sure this is not just an incredible team, but the best culture that we can get to. And it was a really good decision to do it that way. And that’s because A, we enjoyed working there ourselves because of that. But B it meant that we were able to attract some of the best talent from across the world that ultimately helped us build the best products. 

Jon Krohn: Awesome. Yeah. Value creation comes down to the right team. Yeah. And so there’s two big pieces to that that you’re outlining. The first piece is getting the right hires. And then the second piece is having the right culture when they’re on the job and making sure that everyone can be making the big impact that they want to be making, just like you did when you started off creating a startup from the very beginning. Super cool Husayn. This company that you’re working on now, I understand it’s in the creator economy sector, you’re still in stealth, but maybe you can tell us a little bit about what you’re up to, get us excited. 

Husayn Kassai: I can tell you about the problems in that I’ve kind of as you and as almost all of your listeners, we have an appreciation for alternative media, podcasts, YouTube and everything else. I’ve been that kind of a consumer of that content as anyone else has been. Back in 2007, just before I started university, I did a 30 parts series for my high school and college around helping people get into university and sort of pass exams. Recognizing the power of YouTube early on and being a fan since, but then when 2015, 2016 and 2017 were pretty key for me. And then I kind of, at that time, I had this perspective that machine learning AI it’s great, the world’s becoming a better place. It’s all going to be rosy from now on just to exaggerate. And then a series of things happened that were like a blind spot to be like, if I’m looking back, there were blind spots, a big part of that was elections across the world and sort of characters coming in that you would’ve never thought would’ve come in. 

Husayn Kassai: And then because of this deeper understanding of the power of data and machine learning and so on, recognizing that everyone celebrates the social media platforms, but there are certainly downsides. And not only are these downsides bad, but if you just extrapolate of where it’s headed, we’re headed into a pretty dark place. The problem for me ultimately, was one of the media and how it has problems and issues that needs to be resolved. In parallel, I had my own view and thesis around consumption of media can be thought of in two ways. The way I look at it, it’s like the more view count world and then there’s more, what I call like repeatable world. View count world is epitomized by TikTok and behind that Instagram, behind that Facebook typically. Where you want to go to that content it’s typically humor or entertainment. 

Husayn Kassai: We all do it. And it’s good and it’s has it’s place. But broadly speaking, like distraction and switching off. But the other side of the spectrum, we look at repeatable world content, meaning exactly like your podcast, where you may have relatively fewer viewers or listeners still often tens if not hundreds of thousands, but who come back week after week often for many years. And those categories is typically more guru, learning, hobbies and so on. And that they are somewhat underserved. And that these algorithms are heavily disproportionately geared towards view count world and pushing up viral content, especially helping new viewers come as a result of their algorithms. 

Husayn Kassai: But those same algorithms are disproportionately, adversely affecting the repeatable world content. And then the repeatable world content, usually promotion and getting new viewers or listeners comes as a result of true fans telling their friends and peers or other similar content, creators sharing what you’re doing. What we’re doing is essentially helping address that, bringing a more effective monetization and distribution approach to video content creators in particular, in the repeatable world and offering them the tech infrastructure to be able to do that. 

Jon Krohn: Nice. Well, you’re preaching to the choir right here Husayn. 

Husayn Kassai: Right, in our wheelhouse. Absolutely. 

Jon Krohn: Yeah. This is music to my ears because I definitely, I experienced this where we have this podcast, most listened to podcast in the data science industry. And so that’s great, but I think to myself, well, how could we be getting more exposure? And we do get a little bit more exposure through YouTube algorithms. And I think a lot of new exposure comes when people Google Data Science podcast, we come up first. And so then they’re likely to consider us, but we don’t benefit in the same way that viral content does. So because we’re in that second camp you described, the repeat content camp where it’s amazing, we do have listeners like you listener listening right now, who some of you have been around for many years and you listen to every episode. And so it means that every episode it’s really consistent in terms of the number of listeners. You have to look like month over month or quarter over quarter to see changes in listenership, whereas with the algorithms and viral content, stuff that a YouTube algorithm or TikTok algorithm favors, you end up having these singular pieces of content that do incredible. 

Jon Krohn: And yeah, we’re not making content like that. We’re not making viral content. We’re not trying to be like click baity. We’re just trying to give you a consistent, two shows a week and on my YouTube channel, my personal YouTube channel, one video tutorial a week, every single week. And we hope that people stay committed and stay on. And so like you’re saying, word of mouth can be a big thing. People on social media saying, “I just listened to this amazing episode of the Super Data Science podcast. I learned this incredible thing. And I’ve been listening to the show for years. I highly recommend it.” I think that’s one of the primary drivers of how we grow is this word of mouth, as you say. I do, I understand the problem you’re describing very intimately and yeah, I’m excited to see what you’re up to and how it resolves this. Can you give us a little taste or we should just save it until launch time? 

Husayn Kassai: Taste yeah, so we’re just under 20 people. The majority are engineers, coders, programmers working away on our prototype. We are in beta with about [inaudible]- 

Jon Krohn: I love that you’ve been doing this for less than a year and you already have 20 people, mostly developers working on it. Yeah. You couldn’t stand gardening. You needed to do something. That’s- 

Husayn Kassai: That’s what I said. The gardening just sounds so good. But I do, there is a concept of a founder problem fits. And spending 10 years on this, on solving the identity problem, I know yes, to be something that drives you that’s over and above and what’s normal. In an unhealthy way, you have to become a little bit obsessed about solving it. And the problem was so bizarre, so terrible, this identity, how it’s broken and needs to be fixed. And that was like a drive to work and make it happen. That same level of energy and excitement holds for this problem that we as a team are working on. 

Husayn Kassai: And I think that’s really, really important. When I knew, it was quite clear to me that this is a problem that needs to be solved. And I’ve just been working on figuring out how, as opposed to thinking about what industry should come next. It was kind of natural for me after a break that whatever it is ends up being, it would be towards addressing this problem. Because I see it personally as a very fundamental issue, the way that the social media platforms could have an adverse impact on the world. 

Jon Krohn: What was your break, like a three day weekend? 

Husayn Kassai: Well, this is a thing, actually. The reason why I was able to go for like 10 years really well. And I’m sure this gig I’m setting aside like 20 years is because you got to pace it, that this is not a sprint, this really is a marathon. You’d be surprised I actually do a relatively normal five days a week, kind of 8:00 AM to 6:30 PM. It’s not like crazy that you might get these far worse in the city, in super financial institutions. And I spend half a day on one of the weekends and that’s pretty much it. And that sustains me the whole year, I take about two weeks off, but I do it because I genuinely don’t necessarily see it as work. I absolutely love what it is that I do, especially because I work with incredible people and that all helps. That all really, really helps. 

Jon Krohn: Cool. It’s nice to hear that you have the kind of balance. I’ve got to find a way to get more balance and I could see how by not burning the candle at both ends in particular days or particular weeks over the long run, you’re more focused and more productive anyway. And so, yeah, I’ve got to get some more of that balance in my life Husayn. Now that you have this second startup running, do you have a particular process for deciding on what startup opportunity to go with? Or both of these opportunities Onfido with credit and identity, and then now with your stealth startup and the creator economy, were you just intuitively, this is the biggest problem that I could be tackling right now? Or do you have a kind of like an exhaustive process that you go through before you go headfirst into the opportunity? 

Husayn Kassai: As far as biggest problem, it doesn’t necessarily have to be the biggest problem, but has to be what I see as the most important. And it just happens to be fortunate that identity was a global problem and this video content consumption is a global size as well. And it helps when it’s, you’re dealing with a global or something that has a potential to be large on a number of fronts. The first thing is, when you hire people, you can say, “Look, we’re working on a project or a mission that can be global. Therefore, it’s like super impactful.” And therefore you’d get a real kick out of building something that can be big. That’s part of it. The second part of it is, you’re able to get more investments to be able to run fast and win in essence. 

Husayn Kassai: So there are those sides to it, but I do have a very strong appreciation for quality. I know that in the Bay Area, there’s this obsession with release and iterates or run fast break things, whatever. That’s not the way I’ve seen it nor see it, because in this day and age, the consumer pretty much gives you one shot. You got one shot to impress. And if you don’t do that, they don’t care that you’re going to iterate. You really need to do a good job. So I have the reason why we’re kind of taking our time and doing it organically, despite this being a fast paced world is because for me, if it’s slow, steady and done properly, that’s worth a ton more than running fast. 

Husayn Kassai: And the story went fast from a month ago and other stories, really I’m sure hardworking team and everything, but just, they’re not getting the timing and the pacing right. And sometimes it’s undue influence from investors in particular. And ultimately the founders as a decision maker, they should be setting the pace and they should work to align everything else around that. Not have more investors wanting to invest or different components, more team members wanting to join or whatever it may be setting the pace. The pace should be, you should be in the driving seat, not being driven by the need for speed as such. 

Jon Krohn: That’s that cool confidence that you exude that I love so much, that’s it. And that’s it in space right there where you’re like, “Yep, we got to get this product ready. We’re going to be disciplined about working on it consistently, getting it to a point where we can grow it organically. And we’re not going to bow to pressure from the outside to rush this thing and risk not impressing our consumers on that one first shot that we get.” I love that Husayn. 

Husayn Kassai: I do see it in a very binary way. There’s just one other thing, when I’m speaking with and learning from other founders, for me it’s very, very clear cut your pre-product market fit stage, your product market fit stage and then your scale up stage. Where I think it is just each one is a very clear binary progression to the next. Where you’re now getting more pre product market fit stage companies, but investing and growing as like scale up. And you have three, 400 people, but you don’t even have product market fit. That doesn’t help anyone. It doesn’t help you necessarily get that faster, if anything it’s going to slow you down. 

Jon Krohn: Yep. 

Husayn Kassai: But this [inaudible] wasn’t really an option 10 years ago. 10 years ago, you couldn’t do this. You would not get investment unless you had strong product market fits. In some ways, 10 years ago, it was much harder to get investment. But the [inaudible] that led to startups and scale ups coming out meant they were stronger, which is a change in the last 10 years. 

Jon Krohn: Yeah. There’s just so much money flowing now into, in this low interest rate world that we’re in. Though that might change in 2022, the low interest rate world that we’ve been in for over a decade until now means that everybody’s looking for yield. And so you have big hedge funds, big institutional managers that are saying, “Oh, look over there, these VC firms are getting yield above zero. So let’s put a bunch of money over there.” And so I was reading a stat I can’t remember the number, but it was something like 600 billion. Don’t quote me on this listener, but something like $600 billion flowed into machine learning tech startups in 2021. And that was double the amount in 2020. The double, you can quote me on, the exact number, don’t quote me on. But so when there’s that much money flowing in, if you are a VC and you’ve got this giant pool of money that you’re expecting to deploy over a certain time span, you’ve got to put it somewhere. 

Jon Krohn: And so you end up having this situation that you’re describing where pre-product market fit companies are getting scale up level investment, and yeah, it could lead to a rocky few years. Maybe they’ll get lucky, but maybe they won’t. Super cool to get that perspective Husayn. All right. Now we have some idea of how you think about the opportunities that you tackle. We’ve heard a lot about your founding team with Onfido. How Eamon was key for operations and Ruhul was key for technology. Do you have a process or rules of thumb for choosing your co-founding team? 

Husayn Kassai: I’d say a co-founder when I speak with other founders or co-founders thinking through this process, it ought to be seen as again, it’s very much a binary thing. If someone’s a co-founder versus a very senior employee or an incredible team member employee. Whereas the reason why it’s binary is like a co-founder is all in by every definition, meaning they do, there is a trade off to be had. They will have shorter weekends. They will have fewer holidays, they will lose sleep if the money’s going to run out, or if there’s some sort of fire in some part of the organization, they will have to lead by example and the rest of it. And they have to be able to learn and go through a very quick learning process quite rapidly. With all that said, that whole process of whether someone is becomes a co-founder or not, should be thought about very, very carefully. 

Husayn Kassai: Because I do understand that this emphasis around I’m like, I don’t necessarily have a lot of experience myself. Let me bring someone in, they’re only going to join if they get the co-founder title or other titles, like a senior title. There’s this urge to promising stuff and then having it work, but it is such a consequential decision and it should be thought of really, really carefully. Having said that for someone who’s doing it the first time, having just gone through that for 10 years, I would not have been able to do it without my co-founders. And you really need to be very well aware of your skillset and crucially what you are not good at and make sure you have complimentary or one or two or three co-founders. And make sure you’re all super complimentary and very different to each other so that you collectively combine the, make it kind of the full package. With all that said, if you don’t have the full package and if you’re first timer or new, I would recommend a co-founding team just think about it really, really carefully. And unless you have that perfect match, don’t rush into any sort of co-founder relationship, unless you’re absolutely sure because they should not be thought of as like, this is my best senior team member. Let me make them a co-founder because it’s just not going to be like that. 

Jon Krohn: Awesome, great advice. All right. Husayn, you’ve already had this tremendous success scaling up an AI startup, and now you’re onto your second startup. You are generally a wise and worldly person. So I think that this question will be an interesting one for you to answer. In the decade, in the past decades and presumably for the decades to come, there’s been ever cheaper data storage, ever cheaper compute, ever more abundant sensors. There’s been unprecedented interconnectivity between people and more and more people on the planet connected to each other every year. And lots of modeling innovations, data modeling innovations, as well as DevOps, cloud innovations, accessibility to scaling up machine learning models and technology platforms in general. Every year, the pace at which technology changes increases. So given this and given that you’re still relatively early on in your entrepreneurial career, what excites you about the future? 

Husayn Kassai: If anything, it is now that potential to start something and for it to go global exists just that same feeling that me and the other sort of my two co-founders had in 2008 now holds. And in fact holds more than it ever has ever in history. Now there’s just a lot more investments or capital available to be invested in developing things. And whereas, 10, 15 years ago, if it would’ve taken 10 years for a company to be global, now you increasingly have that being a shorter time scale, three years, five years, something can go from zero to essentially having an impact globally in a very, very quick and short pace of time. Your question of what is exciting? That in itself is very exciting and parallel. There can be negative so that naturally, if you don’t have product market fit, then scanning is kind of doesn’t necessarily get you product market fit. There’s that point from earlier, but crucially as and when there are technologies that are deployed and deployed quickly and have a global foot reach in a short amount of time, we all know how regulators can be slow in catching up, it’s just the way it works. It’s, they need to do a much better job of getting much closer to understanding technology. 

Husayn Kassai: I’m sure and hopeful that you increasingly have regulators lifting in on your podcast. So that they’re up to speed with the power of this technology and technologies like this so that they understand there are so many upsides and benefits, but there are downsides. And that it’s our job as a collective, both as whether it’s on the academic side, practitioners, policy makers, or often a bit of all of that to collectively work to put in guides, policies, rules, regulation, so that we make sure we maximize just the good stuff and control the stuff that can adversely affect you. Because when you get to a state where in five or 10 years, we don’t put a control on some of the risks on the adverse things, you can have a backlash. There’s been technological black backlashes in the past, we just celebrated the advances in medicine diagnosis. We’ve just been talking about identity and what I’ve been sharing and all the very many good things that can come as a result, but not be sort of just to remain mindful of the downsides and to work, to minimize those. 

Jon Krohn: Nice, well said. And wise, as I expected. Already Husayn, this has been a tremendous journey getting to hear about your journey from the conception of Onfido, to being a massive global successful company. And then now onwards to your next big thing. Starting to wrap up the episode, do you have a book recommendation for us? 

Husayn Kassai: I’m immersing myself with media and everything media related. I just finished Hate Inc. It’s Matt Taibbi. It just talks a bit about specifically news. It seems like it does a quite fair job criticizing CNN and Fox, and just talking through the issues that there’s a political layer to it, but it’s a bit deeper than that, talking about how and why we, as sort of whether do you look at the US or in the UK 10 years ago, we kind of saw each other all as the same, and now there’s increasing this tribalism problem. This is a very good book that helps go towards explaining some of the reasons why. 

Jon Krohn: Cool. That’s a great recommendation and very on theme for topics we’ve been covering in this episode. I’m sure listeners will reach for that book now. As I mentioned, a number of times on the show, I have enormous respect for you, always come away from conversations with you feeling more focused and refreshed than before we started talking. And yeah, I also feel like you have tons of wisdom to impart all the time. How can listeners follow you after this podcast? How can they continue to stay up to date on your latest thoughts? 

Husayn Kassai: That’s kind of you to say, so I’m on Twitter and my Twitter @HusaynKassai 

Jon Krohn: Nice. We will include that in the show notes, for sure. Thank you, Husayn for taking time out of what is no doubt, your busy life as you get going on this next stealth startup, we can’t wait for that to be launched and hear more about it. Thank you so much for being on the program. And maybe we can catch up on the show again in a few years to hear about how scaling up that startup is coming along. 

Husayn Kassai: Absolutely. It’s great. Great to have been on. 

Jon Krohn: Loved catching up with Husayn on air today. I hope you feel as grounded and inspired from hearing from him as I do whenever I talk to him. In today’s episode, Husayn filled us in on how he conceived Onfido as a firm to disrupt the 150 year old credit footprint industry. How the value creation when scaling a company comes from hiring and retaining a great team. How the US market is often the key to winning the data race if you are developing a data hungry AI platform. How he weighs the most important problem to solve is more important than the biggest problem to solve when deciding on a startup opportunity. How it’s key to develop your AI product at a slow and steady pace in order to find product market fit before accepting tons of venture capital for scaling up. And how a co-founder may be critical for you to show up your own skillset gaps, but you shouldn’t rush into bringing just anyone on as a co-founder, as it’s a big decision reserved only for those who are willing to lose a lot of sleep. 

Jon Krohn: As always, you can get all the show notes, including the transcript for this episode, the video recording, any materials mentioned on the show, the URLs for Husayn’s Twitter profile, as well as my own social media profiles at www.superdatascience.com/577. That’s www.superdatascience.com/577. If you enjoyed this episode, I’d greatly appreciate it if you left a review on your favorite podcasting app or on the Super Data Science YouTube channel. I also encourage you to let me know your thoughts on the episode directly by adding me on LinkedIn or Twitter, and then tagging me in a post about it. 

Jon Krohn: Your feedback is invaluable for helping us shape future episodes of the show. Thanks to my colleagues at Nebula for supporting me while I create content like this Super Data Science episode for you. And thanks of course, to Ivana Zibert, Mario Pombo, Serg Masis, Sylvia Ogweng, and Kirill Eremenko on the Super Data Science team for managing, editing, researching, summarizing, and producing another inspiring episode for us today. Keep on rocking it out there folks, and I’m looking forward to enjoying another round of the Super Data Science podcast with you very soon.