I initially joined Flow Labs before it was even called Flow Labs, back when it was known as Flo AI.
In the late summer of 2017 I was a 3rd year PhD student at McMaster University researching deep reinforcement learning for adaptive traffic signal control when I received an email out of the blue from a guy named Jatish Patel. He wrote that he had read some of my research that I had published on arXiv
and was interested to talk to me. This was the first time somebody had contacted me about a paper I had written so I was more than happy to talk with him. Over the next year we talked on the phone twice about traffic signal control and why, despite many attempts, adaptive systems had never achieved widespread success. By late 2018 I was finishing up my PhD and Jatish had established a relationship with the Utah Department of Transportation (UDOT), which led to him offering me a contract to help optimize some of their traffic signal timings. We delivered optimized traffic signal timings to UDOOT in early spring 2019 and waited with bated breath for their deployment results. Fortunately, our hard work paid off and the results were better than we could have hoped for, with all of the key performance metrics changing favorably.
However, this was a bittersweet moment for me, as it was the end of the project and Jatish had long made it clear that even if we were successful, he could not start a company without sufficient time to raise investment. This left me in a contradictory state of being pleased with our success but depressed that it had to come to an end. Luckily, by late 2019 Jatish had raised sufficient investment to start Flow Labs and I was able to be a member of the company’s founding engineering team. I initially joined Flow Labs for a few reasons. First, it provided me the opportunity to take ideas out of the lab (where I spent my time in graduate school) and test them in the real-world. I suspect any good grad student at some point will wonder if the theoretical knowledge they are taught actually works when the rubber hits the road. Almost all of my graduate research employed traffic simulations.
It had also been a great experience working together with Jatish on Flow Lab’s first proof-of-concept. The problem was technologically interesting, we were doing something meaningful that positively impacted society, and the initial results were very promising.
The last reason was that I wanted to challenge myself. My first real job after completing my PhD was working as a Senior Research Associate at the University of Toronto, which from all sides was a great job. I was involved with almost a dozen technical projects which included many brilliant students and professors. However, at this point in my life I had been in academia over a decade and had known nothing else.
Working at the University of Toronto I could easily see my future unfold in front of me. I started asking myself if this is the career path I wanted to stay on. I decided I would regret not taking the opportunity to gain industry experience, to test if the theoretical translated into the practical, so I left academia and joined Flow Labs. Additionally, I was curious if the pop culture sentiment surrounding tech startups was fact or fiction (it’s a big of both). I also thought I would have more agency in a startup than at a university or large corporation, which was alluring to me.