From PhD Student to Machine Learning Entrepreneur: SuperAnnotate CEO Tigran Petrosyan (Part 2)

Sramana Mitra: When you applied to the Berkeley accelerator, how much time passed between your $100,00 funding to the accelerator?

Tigran Petrosyan: We got accepted to Berkeley after two months and then into the program after four months.

Sramana Mitra: How long were you in the accelerator?

Tigran Petrosyan: It was six months. I’m not sure how long it is right now. It’s one of those places in the Bay Area with a terrific view of the Bay.

Sramana Mitra: What milestones did you achieve during that year?

Tigran Petrosyan: Initially, you’re just doing things. You don’t think about what is right and what is wrong. We built a 40-people team in terms of engineering.

Sramana Mitra: All in Armenia?

Tigran Petrosyan: The majority was in Armenia. We already had our first two paying clients.

Sramana Mitra: How did you get the paying clients and where were they?

Tigran Petrosyan: The accelerator gave us some visibility. Through the network, we managed to meet some companies and then eventually start working with them.

Sramana Mitra: What kind of companies were these?

Tigran Petrosyan: One of the companies was building autonomous trucks.

Sramana Mitra: Not medical imaging diagnostics.

Tigran Petrosyan: Not in the beginning. It was more about how we use our technology in segmentation. It didn’t have to be in medical imaging. This is where we started – autonomous driving.

Sramana Mitra: How big was this contract?

Tigran Petrosyan: It started small in the beginning – a few thousand per month. It grew to a few tens of thousands per month.

Sramana Mitra: In terms of product positioning, what did you learn?

Tigran Petrosyan: Basically, working with early clients teaches you much more than any other theory. We had our tech but realized that there’s much more than our technology. The problem we needed to solve was building our product while working with this customer.

Sramana Mitra: Was your technology now positioned for the autonomous driving market?

Tigran Petrosyan: At that time, it was being built for the autonomous driving space. It was not just one type of labeling task but multiple types. It’s not just segmenting images but also putting boxes in images. It was much more than the initial technology.

Sramana Mitra: Did you get other autonomous vehicle customers?

Tigran Petrosyan: Eventually, we started to get other customers because we were not mature at that time to go for other autonomous driving customers. Autonomous driving was flooded with so many applications. Just because we had a connection doesn’t mean that it is the right go-to-market strategy.

Sramana Mitra: This is where I wanted to counter with you when you said you learn more from customers than theoretical analysis. I would argue that some theoretical analysis is necessary before you decide which customers to go after.

Tigran Petrosyan: Absolutely. This is where I appreciate the accelerator program. We were exposed to so many talented people who were experts in the industry. There were very famous names including even the early founders of Tesla. Those conversations were important for us to shape our understanding of where our technology can be most used.

As we build, that helps us better understand what customers want because having the tech is basically 10% of everything. You can have the best tech but if you don’t have the right team to execute and the right way of serving customers, you’ll not succeed.

This segment is part 2 in the series : From PhD Student to Machine Learning Entrepreneur: SuperAnnotate CEO Tigran Petrosyan
1 2 3 4 5