Kris, thank you so much for sitting down with me today. Could you introduce yourself to our readers?
Absolutely! I originally come from an academic background and finished my PhD in AI in 2006. I was very fortunate to have a lot of freedom with my thesis and over the first year, I chose a topic which really fascinates me – which is How do children learn language? and building computational models of it. This allowed me to look at research in cognitive science, linguistics and neuroscience which are fun areas that appear in AI. I could spend a few years geeking out on a topic that (at that time) I thought had no real money or commercial applications behind it. It was really research for adding to the body of knowledge in the world, so that was lots of fun!
When I finished, I had to think about what to do next. Do I stay in academia or do I go into industry? It was a tough decision. When you leave academia, you’ve got a certain window of time where you can realistically go back into it. In academia, the world is set up around your academic publications. So if you stop publishing, it’s like the currency of your world dries up and you can’t go back into it so easily. So a lot of people go out of it to see what it’s like and then step back in. I did it three times as I did three different post-doctoral positions. They were all fantastic. Each one, I learned something completely different from.
First off I moved to France Telecom, or Orange as it was called by then. I did 18 months learning about personalised information retrieval; so personalised search, and information filtering, which most people call recommender systems. I did a lot of interesting work there but I got a little frustrated because what I was producing never got out as an actual product. It’s a common problem in research where work ends up on the shelf.
Then I changed and moved into a different Postdoc and went and worked for a French company called CEA (the French Alternative Energies and Atomic Energy Commission), where my role had nothing to do with atomic energy. I was working in what was rumoured to have been a ‘James Bond’ department at some point in time but I can’t talk too much about that! But, let’s just say, a lot of things that exist now, these companies were working on many decades ago. What was seen as sci-fi back then, they were looking to crack these problems. Can you imagine giving a device that looked like a tablet from nowadays to someone, so they could automatically scan documents written in another language and translate it for them, all whilst connecting to the internet to tell them what it meant. So really fun and cool stuff!
Then I went to do my third Postdoc, where I moved back to the UK and joined the University of Manchester and worked in the national centre of text-mining for 6 months. That was my last position in academia.
After the University of Manchester, I discovered a company called Mendeley, who were making tools for researchers. It was an early London tech startup, which excited me because even though the research I was doing in academia was exciting for me, I never really got it into mass public hands. It would be the case, far too often, that great work from my colleagues would end up sitting on the shelf – which is very sad and very frustrating. I also wanted to try out a smaller company as I’d worked for both medium and large companies, but never a startup, so I wanted that experience. I found my home when I went there, it was a fantastic experience. I stayed with Mendeley for several years, specialising in mostly recommender systems technology, although working on a wide range of data science projects.
We were then acquired in 2013, where I went on to become the Head of Data Science for Elsevier (who acquired Mendeley) – they’re the largest academic scientific publisher in the world. It was a great privilege to be a part of projects that were more substantial and significant, that would reach more people with actual products. The flagship product that I am most proud of from my time there is Mendeley Suggest, which is like Netflix or Spotify for academic research papers.
So, a bit of jumping around in my career, but the thread that holds it all together is information systems. How does information get in our heads (i.e. language learning), get out (i.e. language production), how do we guess other people’s preferences, and how do we help them find things?
You moved from Elsevier to TUI late last year (who are quite different!) – what made you make such a big step in industry?
TUI was another great opportunity to take as travel and leisure is one of the top industries that’s been found to benefit most from AI and data science. There’s a huge amount of potential to improve people’s experience while on holiday. TUI is the largest travel and leisure company in the world, so it was great to see it from their perspective and apply the types of technics that I’d learnt over the years to the problems they had.
So, why create Smart Tribe? Where did this all come from?
I was reflecting on my own career path and when I was looking at how difficult it was to decide on whether I wanted to be in either academia or industry, I realised that there wasn’t a single tool or place to go to that will help with that decision. So, my problem was that I had read, learned and produced some nice research but didn’t know to find companies that would make the best use of my skills. Researchers tend to focus on their career in terms of academic publications as this is how you funding for future research but industry does not, there’s not a lot of overlap between the two. Academia and industry basically don’t talk the same language.
Industry is often (and rightly so) very discreet and cautious about telling the world the types of problems they’re trying to solve. Many times, when you read a company’s job advert, it often describes a general set of skills that they are looking for in a candidate. It doesn’t reveal the detailed specific problems as this would give away a competitive advantage if other companies were to read it. At the same time, from an outsider’s perspective, it’s hard to get noticed and showcase your skills as they are all in research publications. There’s not really a place where you can go to showcase your skills in a language that is accessible for industry to pick up.
Having experienced that myself, when I met my Co-Founder (Beatrice Zatorska), we instantly clicked. She has been working in commercialising technology for the past 20 years, and most recently in a management consultancy role. She started in the defence sector and pretty much moved into every other sector that’s ever existed along the way. She would go into companies (usually under NDAs) where they would explain their problems and needs and she would go out and find the right talent and technologies that can help solve it. Sometimes it would be finding tech they could buy, sometimes it would be finding an expert who could come in and help them, as well as pretty much everything in-between. She would sometimes take research from the ideas stage all the way through to marketing it and taking it out to the real world.
So, she’d been working on solving a problem that I had experienced myself. The problem of specialised people finding it hard to find companies that needed their skills and vice versa. So, we both met each other through Antler, which is a startup generator and early stage VC. We started discussing with each other our different experiences and realised we were both obsessed with solving this problem. Even when we speak now (because she comes very much from an industry perspective) the two of us sometimes have challenges talking about the same thing together because we have different ways of explaining it and different words to describe it. We’re like a very small-scale example of what’s happening on a much larger scale, with two worlds who don’t speak the same language – but we really like each other so make the effort! We’re figuring out how we can put a solution in place where we can help others get over these language barriers.
So Smart Tribe is a place where you can go as a researcher and describe your research (or research you know well) in plain English. Companies can search through those plain English descriptions (which we call Blitzcards) and can find solutions to their problems, all in private. It allows companies to search for the solution to their problem in confidence and from those Blitzcards, they can reach out to people with that expertise or technologies.
That’s amazing. How does the platform work?
We’ve been speaking over the years (and really focused on the past couple of months) to as many researchers, companies and the players in between these two worlds. There’s a particular office called ‘Technology Transfer Office’ that many universities and institutes have attached to them. Their reason for being is to get the research out of the university and to get it commercialised and into the real world, and some are doing a really good job at it. Some countries have really perfected this method, but for many others, it’s tough because they don’t have enough people working in the tech transfer office to represent all the different research in their institute and keep on top of it. Because industry has its need for privacy, they have to spend quite a bit of time building these relationships to know who needs what.
We’ve been speaking with a lot of technology transfer officers to find out their needs, pain points and what works and this is what we’ve found to be the biggest problem – not being able to keep on top of all that research and not knowing all of the demand. This tends to lead to a lot of localised transfers, building really good relationships with geographically close companies. So, we want a more globalised solution, which would allow a company in Australia (who needs a certain technology) to find the solution in somewhere like Morocco. Smart Tribe uses machine learning to help companies find tech relevant to their needs and to connect to specialists who can help solve their problems.
What made you finally take the jump at starting Smart Tribe?
I really need to solve this problem!
I get so excited about this. I love this problem and I think it’s important that people work on what they love doing. I don’t have a more eloquent answer to that, unfortunately.
What made you go with Antler, over other larger cohorts?
Antler is an interesting startup generator and early-stage VC, that brings together groups of people who want to create startups in tech. It’s run by a strong team of experienced entrepreneurs and professionals. At times, it feels like being in an episode of The Apprentice, at other times Dragons’ Den. They put your feet up to the fire on the important issues and want to see you succeed. My group is based in London and has around 75 people in it.
They take time to interview everyone and it’s a very selective process. They have done their homework about what makes a successful startup and what the problem areas are to avoid. They give masterclasses that are fantastic and help us to think about many of the characteristics that are important in setting up a startup. They focus very much on the founding team for each startup being strong with complementary skills and shared values. I’m very glad to have chosen Antler as I have found Beatrice through it.
The big problem you’re trying to solve with Smart Tribe is the communication between academia and industry. What issues do you think your work will have the most impact on?
One of the big problems between the two different worlds is that they don’t talk the same language, and there are two aspects to that.
One is that academic literature is loaded with specialised terminology which often makes it hard to read through a paper. You get trained for years to be able to read research, as well as produce and disseminate it. The second is a problem with the language side of things. It’s not just an issue across from industry to academia. Even within academia, different fields of expertise find it difficult to read each other’s work.
The biggest pain point from the industry side is you can’t easily access and search through these academic solutions as they’re written in a different language. We’ve had a lot of interest from companies as they’re after a search tool, where they input their problem in plain English and find the right tech and talent that solves your problem. So, we thought this may be a bit of a moonshot but we’re going to try and build this!
The main issue around the two worlds is that there are different incentives for both of them. From the industry side, you want a solution as soon as possible. You often don’t need a novel solution, but if you don’t know it exists it’s hard for you to solve your problem. As for academia, you’re incentivised and your career grows based on your publications list. Academic articles need to have a certain level of novelty to them, meaning you can’t produce a result that’s already been published – so there’s always this incentive to do something new. Those two things are misaligned. Companies need solutions quickly and whereas academics need to creating novel solutions, which often don’t come quick.
What we hope to do through Smart Tribe is we’re asking researchers who have all the expertise and knowledge to write these plain English translations (Blitzcards) of new and existing academic work. We’re asking academics to take a step towards industry. This is something, for most people, that usually happens at a much later stage, where you have to explain it in a lot of detail.
When it comes to the recruitment side when finding top talent, we’ve been getting a lot of feedback from companies saying when they’re interviewing candidates, they’re not just after very technically talented people, but are looking for people with business acumen who can take the work they’ve done and can apply it in such a way that it will benefit the company they’re working for. The traditional job description and CV approach doesn’t do a very good job at this at the moment, meaning that you need to get through the screening process before you can explain how your work solves their problems. We want to move that to the initial conversations.
Was Smart Tribe founded on the pain points you had whilst moving into industry?
Certainly! When I was finishing my PhD and I didn’t know what my options were. I would search different job listings and at the time the work I was doing was not as ‘hot’ as it is now. Finishing a PhD in AI now would be a lot more popular, but at that time, there weren’t obvious applications for it. I wasn’t being approached by recruiters and people weren’t noticing what I was doing. I think that’s true for a lot of early-stage researchers until you need to build a name for yourself. There’s nowhere you can really go and search for a company that can use these skills – it’s tremendously hard. The job descriptions had abstracted away from the real problem enough that many of them didn’t look very enticing for me. It made it very hard to match very specialised skills with an industry that needed them. I believe that there’s often a better possible match than the ones we currently see.
It’s worth noting that going into industry isn’t for everyone, so creating an academic partnership is always an option. Only 20% (on average) of academics tend to be interested in commercialising the academic work they’re doing. Most academics do it to advance the world’s body of knowledge on subjects, which is a noble profession and money can get in the way of it. If academics wanted to be in industry, they would. For the few that are interested in going into industry, it’s not obvious how to find the companies that will make the best use of your skills and that you’ll have a job that you’ll adore and geek out in.
Now you’ve started Smart Tribe, how have you found it? Going back and forth from academia to industry and now becoming a founder must be very interesting.
Yeah, I love it!
I haven’t spoken to as many people about their problems and how we can help solve those problems in such a condensed space of time as I have in the past two weeks. It’s let a lot of lightbulbs flash and it’s been a really exciting design process to follow. We follow an ethnographic process where we’re trying to understand what people’s current problems are and how they go around addressing them at the moment, all whilst listening out to the problems that evoke real emotions from them. We then know that we’re trying to solve a problem that’s worth building a company for.
This has been a fantastic experience. I’ve loved talking to researchers and companies, as well as everyone in between, who are often very passionate about this problem as they care about research going somewhere. Now we’re giving them a platform where they can market that.
Finally, any advice for Data Scientists and Researchers in taking that leap and finding their own dream and passion?
For me, what helps the most is if you tackle a problem that you’re very passionate about. The solution will change, probably every day to start with and will continue to for a couple of years. Similar to if anyone’s ever done an MSc or PhD before, you need to have a lot of self-motivation to reach the end of it and if you’re not really into the problem you’re solving and not into what you’re doing, then you may give up along the way. It’s completely normal as it’s a very hard journey and I think it applies a lot to startups, as in startups you have to be everything and you have to wear every hat. It’s going to be a much harder experience if you haven’t fallen in love with that problem.
That would be my main bit of advice, as the technology changes you will always learn new skills and things. That problem needs to be something you wake up in the morning and think ‘I have to solve this, the world needs that solution’. I think the solutions to a lot of the world’s big problems like world hunger and the climate crisis can be addressed with impact technology and make a difference, a lot of answers are in academia already and the answers aren’t getting out there.
That’s my personal motivation. I adore the problem I’m working on – I want to solve it. What doors can we open for this flow of knowledge to go faster?
Kris, thank you for taking the time out and speaking with me about yourself and what you’re doing at Smart Tribe. I look forward to following your journey!