From PhD Student to Machine Learning Entrepreneur: SuperAnnotate CEO Tigran Petrosyan (Part 3)

Sramana Mitra: What emerged as your target market?

Tigran Petrosyan: This is a good question. Machine learning can be applied to all kinds of industries. Initially, our target was computer vision. Basically, those who use images to get intelligence out of images. This was our niche.

We started to build an engine to attract clients. It went through a few iterations and another six months before we realized that what we were building is much more than an annotation platform. There’s a big pain about the platform itself as a standalone software solution. We released that in early 2020.

Sramana Mitra: Targeting what market?

Tigran Petrosyan: When you’re thinking about markets, companies are building machine learning.

Sramana Mitra: So your customers were other machine learning companies who were building models and you provided the data and annotation infrastructure.

Tigran Petrosyan: Yes.

Sramana Mitra: So you went into machine learning operations. You became part of the toolkit that ML companies needed.

Tigran Petrosyan: Exactly.

Sramana Mitra: There’s a services component to it?

Tigran Petrosyan: Yes, that’s a very important part here. Human feedback shapes those models. Finding the right teams that can build those datasets correctly is critical. We built a marketplace of teams across the world who are building those annotations.

We knew which team is strong in which area. How many people do they have? What kind of security measures do they have? We match them with the right client and we manage those teams directly with our service operations team.

Sramana Mitra: You have a platform business that has a SaaS business model?

Tigran Petrosyan: Yes.

Sramana Mitra: These teams that do human annotation work, do they get the projects or do you take the projects under your company, and then you funnel the project to the right teams?

Tigran Petrosyan: We handle the direct customer relationship. We basically become the client of the partners.

Sramana Mitra: How many such teams are we talking about?

Tigran Petrosyan: We have worked and vetted over 200 teams. We work with about 50 to 70 teams on a regular basis.

Sramana Mitra: What sized deals are we talking about?

Tigran Petrosyan: Teams can have 100 to thousands of labelers.

Sramana Mitra: What countries are they from?

Tigran Petrosyan: They’re mostly from Asia – Philippines, Bangladesh, and India. Also, we have teams in Europe and the US. Sometimes, the customers need the teams to be in a specific geography.

Sramana Mitra: How many machine learning companies are you working with?

Tigran Petrosyan: We are working with over 200 companies. Many of them are Fortune 100 companies.

Sramana Mitra: Let’s go back to the point where you made this decision to become an MLOps company. What was the go-to-market strategy?

Tigran Petrosyan: Our customers are those teams that are building ML applications. The important part is how you get in front of them so they know about you. One of the key things is writing very professional content that, any time they want to learn about machine learning, they find us through our content and webinars.

Sramana Mitra: Where do you publish this content?

Tigran Petrosyan: We publish on our website and also share on LinkedIn and other social media. Of course, there’s SEO.

Sramana Mitra: All your leads come from content marketing?

Tigran Petrosyan: Initially, yes. Eventually, we started to develop our own sales and business development to reach out to customers. Inbound is still the biggest part.

Sramana Mitra: Nothing beats inbound.

Tigran Petrosyan: The exciting part is word of mouth. Whoever worked with us told their friends and other companies. We are excited about that part.

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