Building Version Control for Science: A Coffee Chat with Bogdan Knezevic, CEO and Co-Founder of Kaleidoscope
Bogdan dives into what motivates him to compete, why he cares about scientific collaboration, and how he combines both through Kaleidoscope.
“We want to pave the way to a world where all of science is reproducible and all of scientist time is spent meaningfully”
At Hummingbird we obsess about finding exceptional talent. The founding team isn’t a checkbox for us; it is the whole box.
Earlier this year, we partnered with Kaleidoscope with the vision to build version control for science. The eclectic energy of its three co-founders Bogdan Knezevic, Ahmed Elnaiem, and David Yen was palpable from our first conversation. I spoke with its CEO and Founder, Bogdan Knezevic, to learn what motivates him to compete, why he cares about scientific collaboration, and how he combines both through Kaleidoscope.
Before we dive in, let's take a step back as Bogdan’s story is important. When he was three years old, Bogdan moved from Serbia to Canada to escape the civil war. He witnessed his family’s incredible effort to start from scratch in a new country, which instilled in him a ruthless worth ethic. Bogdan was swimming competitively by age nine, which led him to compete at multiple World Championships. He went on to study Neuroscience for his Bachelor’s and subsequently completed his PhD in Computational Genomics at Oxford as a Rhodes Scholar. Before starting Kaleidoscope with his high school friend Ahmed and Ahmed’s colleague David, he was a founding member of Entrepreneur First’s Toronto chapter.
[Pablo] You’ve mentioned how critical swimming was to who you are today. What lessons did you learn by swimming competitively and studying at the same time?
[Bogdan] Choosing to pursue swimming as seriously as I did shaped a lot of my foundation and value system: ambitious long-term goal-setting, grounded in manageable short-term goals; relentless time-management; learning from failure; the power of team cohesion and shared success; thriving in challenging environments. There’s a lot I can add to this list.
There is also an attitude that I’ve tried embracing: we only grow through challenges and by facing the unknown. I picked my undergraduate major in Neuroscience because it was full of unknowns. When it came to my PhD research area, I chose computational genomics and drug discovery, a completely different field to my undergraduate degree. I chose this path because I wanted the challenge of diving into a completely new field, regardless of the fact that I didn’t have the necessary background. I had never written nor read a line of code prior to starting grad school, while seemingly everyone in the lab around me had previously done a Masters in the field, so you can imagine the discomfort here.
You certainly didn’t make your life easy for yourself. What is driving you to seek the unknown, compete to excel and work in challenging situations?
I’m motivated by my own family’s history; we were immigrants from Serbia, so I witnessed the tenacious grind that my family was putting in, in order to start from scratch and get situated in a new country. This was instilled in me from a young age and I’m driven to carry this work ethic forward and make the most of the opportunities created by my family’s collective sacrifices.
Kaleidoscope’s vision is definitely challenging. Where did this ambitious goal come from?
We started Kaleidoscope to pave the way to a world where all of science is reproducible and all of scientist time is spent meaningfully. It was really born out of frustration at the state of software available to scientists. I used to work as a drug discovery scientist within a highly collaborative environment as part of the Structural Genomics Consortium when I was based in Oxford, and yet a lot of the work felt extremely inefficient and clunky. There was a massive delta between the quality of exciting scientific advances and the quality of software underpinning a lot of that work.
For example, society has made leaps and bounds in the ability to collect and analyze vast, multi-omic datasets; on the other hand, there is a real struggle to meaningfully track this data, along with the context and decisions associated with it — lots of people and teams are still trying to do this in spreadsheets. As a result, important data gets lost, expensive experiments are forgotten or repeated, valuable scientific time is wasted, and so on.
It’s hard to accept that this is the current situation.
Yes, it’s shocking. I remember finding out at the lab that there was a dataset that cost around $100,000 to collect, process, and analyze. The existence of this dataset wasn’t visibly recorded anywhere as it was likely buried in an inventory spreadsheet, it’s relevance to multiple projects wasn’t mapped as it was likely mentioned somewhere in a slide deck or email, and its importance to the rest of the team’s science was not codified in any way. When the collaborator who generated all of that data moved labs, their knowledge of this context was lost, and the dataset went unused, despite the significant time and money that was spent on it. And that’s just one top-of-mind anecdotal example; this happens all the time, at a large scale. It happens across most labs, biotechs, pharma companies and really anyone that deals with R&D.
Experiencing this first hand was very motivating — and not just regarding this inefficiency in the work being done, but also this nagging question of ‘what if?’ What if scientific teams had a better arsenal of software for doing work? What if organizations didn’t have to reinvent the wheel and solve the same internal data or workflow problems from scratch? It felt like there was this real opportunity to accelerate major technological advances in the world by focusing on giving scientists superpowers for doing science better.
How is Kaleidoscope answering these nagging questions?
We’re really tackling three things: collaboration, reproducibility and scalability. We enable collaboration by bringing teams together more effectively and help them leverage their collective knowledge and expertise. We increase scientific reproducibility by ensuring that all decision-making has been codified along the way, which is especially challenging in R&D where you don’t know ahead of time which experiment or project will yield an exciting result. We help manage a growing body of outputs, not just at the earliest stages of work, but also when it’s time to scale the science, which might mean running hundreds or thousands of experiments a week, and we help teams benefit from the compounding gains of org-wide knowledge.
It’s eye-opening how often teams are recreating the wheel when it comes to internal tooling for doing these things; not only is this highly expensive, multiple engineering salaries and multiple years of work, but they also most often end up with something that simply breaks at a certain scale.
Where do you think the biggest opportunities lie in software for bio?
I think the number of problems is wide, with lots of opportunities for multiple companies to be built in the space. This is based on the breadth of things that we hear customers talk about which are unrelated, including areas such as data visualization, inventory management, data infrastructure, and collaboration. This is also based on the fact that specialized fields like bio have unique needs that often aren’t addressed by horizontal tools. When you couple this with the growing demand for solutions and with the fact that so much expensive, suboptimal internal building is happening, it becomes easy to imagine a future where many companies are built around these individual problems.
Some people might argue that existing software-for-bio companies are already trying to tackle a number of these problems and that the space is not yet big enough to support multiple new players, but I would disagree. The problems that many of the early players were built around are fundamentally very different from many of the opportunities that exist today, and the field of bio is in its nascent stages and starting to grow quickly.
Identifying solutions for these types of problems requires a deep understanding of scientific workflows, but it also requires the ability to think laterally about them. How do you balance challenging the status quo with staying grounded in your understanding of your target customer?
The problem we’re working on is not just a scientific one, or an engineering one, or a product one. It’s all three. Having each of us founders come in with deep experience in one of those areas means that we benefit from both breadth and depth of experience relevant to the problem at hand. I also think that having fresh perspective can be a big advantage – for instance, while it’s really important that I, as a former scientist, can explain how something in a scientific workflow works and why it matters, it’s also really powerful to have someone like Ahmed or David push back and ask “does this have to be this way?” What this lets us do as a team is strike a healthy balance between ambitiously challenging the status quo, and remaining grounded in our understanding of the fundamentals behind our target customer. And I think this balance extends beyond us three as co-founders. Our broader team brings together diverse people with deep experience in engineering, product, and business, while also preserving that scientific context.
What made you want to start a company together?
We’ve known each other for a very long time — I went to high school with Ahmed, and he and David worked together at Socratic from the beginning through to the Google acquisition. Funnily enough, I was introduced to David by Ahmed several years ago, by him saying “you two should meet because I know you’ll get along and because the three of us should build a company together”. David hosted me in New York before we ever knew each other well, which helped us build that necessary trust. Through spending time together over the years, we learned that we shared similar values which we expect from anyone that we bring into the company today.
It is also the fact that we were all turning down potentially lucrative jobs, and coming back to the idea that we wanted to build Kaleidoscope, which really solidified our commitment to starting the company together.
And then when you look at our backgrounds, it comes together nicely, especially for the space we’re building in: Ahmed is a deeply experienced software engineer, David comes from a product and design background and I’m the scientist of the three of us, having personally suffered through a lot of the problems we’re working on solving today.
One thing that was noticeable from the start is how selective and careful you were about every hire you brought on. Maja, the Head of Ops, said that you told her you’d be working together when you started a company, several years ago. What are the traits that make you want to pull someone into the company?
The first few employees we brought on were all people we knew and trusted, like Maja. We wanted to bring in people who could take ownership of work and who we had deep conviction in to deliver high quality output. David, Ahmed, and I have worked with some extremely talented engineers and operators in the past and we knew that, at some point, we would want to bring them onboard.
It’s hard to generalize but there are certain attributes that I’ve seen in people that make me want to work with them. I would say the three of us look for humility. I think it is dangerous to think that one cannot fail, and humility can make you aware of those potential blindspots in the company. There are also individual characteristics that might make someone special. Maja, for example, won’t give up until she’s found a way to crack a problem, and she shows a deep interest in the problem space we’re working on.
For the technical aspect, we know that adding engineers doesn’t grow your engineering capacity linearly; it's a quickly diminishing curve. Going from one to two engineers doesn’t 2x your engineering productivity; it’s probably more like 1.3 to 1.5x. So the question for us is: how do you de-risk an engineering hire to decrease the amount of downtime for the rest of the engineering team to bring new people up to speed? We like hiring engineers with strong fundamentals, ones with computer science or biological/chemical sciences alike. Accompanied with start-up experience when possible. We also think it is important to bring in a blend of people who maybe have an understanding of the problem more intimately. In fact, half the team at Kaleidoscope have PhDs and have experienced the problem we’re solving first-hand.
But I also think trust goes both ways. We want to hire people that trust us and know that we will do everything we can to build something big here. They're joining because they believe that we can do something incredible and because they trust that, as we start doing things that are incredible, we will pay that back to them. We want to look after our employees, and we will find ways to give them increasingly more reasons to believe that they made the right decision by joining Kaleidoscope over the numerous other options they had.
Thanks Bogdan, that’s super informative. Anything else you'd like to add?
I’d maybe like to double down on my belief in the importance of a collective effort or approach to pushing science forward. There really is a tremendous opportunity to elevate how work is done and how many people we can impact in the world with various scientific advances. The sooner we as a community embrace this as a ‘team problem’, the faster we will get there. There will still be lots of competition, this is a healthy thing, but until we truly embrace certain mentalities as a field, such as better weighing of build versus buy, more open public APIs and less platform lock-in, etc., we’ll always be rate-limited in getting the most important stuff done.
Closing Thoughts
We think the way science is done needs updating if we want to increase scale and productivity. We hope that by bringing repeatability, reliability and improved collaboration, Kaleidoscope will enable larger outcomes in biology.
If you’re interested in learning more about Kaleidoscope, message Bogdan here.
If you’re a founder, scientist or employee in biotech, message Pablo here.