Thought Leaders in Artificial Intelligence: Comet ML CEO Gideon Mendels (Part 5)

Sramana Mitra: Of the millions of people who are pursuing a machine learning career as a citizen data scientist track, data scientist track, or a developer track, in what time frame will that impact industry?

Gideon Mendels: It’s a gradual change. It’s already impacting the industry. When I studied computer science, there weren’t any machine learning courses in the curriculum. You can take one as optional. Today it’s core. Even if they start in a software engineering role, having that knowledge is significant. The impact started five to seven years ago.

The other side of it is, when you think about a lot of the technical executives in these companies, today it’s people who started their career in software engineering and might not have exposure to ML. As time goes by, more and more executives do have that experience. That’s what’s exciting for me.

The biggest gap is not tooling or education. It’s figuring out how to align what’s important to what’s possible. The line of business owner knows what could be impactful but might not be able to say that this is solvable with ML. The data scientist knows it, but they might not know how to bring that together. The most successful teams are the ones who can figure out how to close that gap. That’s what I’m excited about.

Sramana Mitra: What do you think is the timeline for that?

Gideon Mendels: We’ll probably see them in the next three to seven years.

Sramana Mitra: Very cool. I did three startups, two of which were AI startups in the 90s. It was way too early. It’s interesting that it took so long for machine learning and AI to hit industry in any reasonable form. Now we’re seeing the next level of adoption.

Gideon Mendels: We started five years ago. Now we operate in MLOps. The term didn’t exist when we started.

Sramana Mitra: We’ve done quite a few MLOps companies. That didn’t exist at all.

Gideon Mendels: When we started the company and started talking about experiment tracking, people say why can’t I just use GitHub for that. Some investors even said, “I don’t think this is a big enough market.” I’ve heard that a lot back in 2017.

Sramana Mitra: Is there anything else you want to discuss?

Gideon Mendels: I would say if your company is working in machine learning and not seeing the ROI you expected, there are a lot of levers that impact that. Part of it is investing in the right methodology of how to run these projects. Having a proper stack and tool chain to be successful is a game changer.

Sramana Mitra: One last question. Within MLOps, what are some open problems that you would encourage people to go and solve?

Gideon Mendels: It’s a hard one to answer mostly because of how fast this industry is moving. Today’s problems will likely not be tomorrow’s problems. One of the things I’m excited about is, there is going to be more discussion around AI regulation. I have no doubt someone will build a successful company building some of these issues. We don’t know exactly what it’ll look like, but how do you manage this communication with the regulator? How do you make sure that you’re adhering to these rules?

Sramana Mitra: Thank you for your time.

This segment is part 5 in the series : Thought Leaders in Artificial Intelligence: Comet ML CEO Gideon Mendels
1 2 3 4 5