Q&A with Gleb Evstropov, Director of Engineering

Could you share a bit about your career path before joining Nebius? What led you here, and what made it feel like the right move?

I come from a tech background rooted in competitive programming. I was doing it at a very high level through school and university: World Finals, International Olympiad in Informatics (IOI), that kind of thing. But I also loved being on the other side — teaching high school kids, preparing competitions, coming up with new tasks. I did that for years alongside competing.

After university I started a PhD in computer science but dropped it after three years. By then I knew theoretical research wasn’t for me; I was drawn to practical work. I’d done five internships by that point, including two at Meta and two at Google, with the latter ones focused on ML. When it came to choosing a full-time role, I joined a major tech company back home in 2018. Google offered me a position in Zurich, but at the time staying in Moscow made sense.

By May 2022 I was living abroad, and was offered the chance to help build the team that would run compute and network services for a new cloud. Nebius at that time was still just an idea. I joined officially in November 2022.

You’re a two-time International Collegiate Programming Contest (ICPC) World Finals Gold Medalist and an IOI Silver Medalist. What did those years teach you?

When you do something seriously for eight years, thirty hours a week, it shapes you. There’s no way around it. And I didn’t stop at eight years; I spent another seven teaching and preparing competitions, so in total I’ve been deeply engaged with algorithms.

What it gives you are two things. First, it trains your brain: abstract thinking, the ability to hold multiple objects in your mind simultaneously, fast logical reasoning. It’s like making your cognitive window larger. Second, you build an enormous library of algorithms and approaches. Learning a lot of algorithms is like learning a lot of recipes for cooking. If you know one recipe, you know one recipe. If you know three, you know three. But if you know two hundred, it means you’ve actually learned a lot of basic techniques. Things you can now combine to produce your own new recipes.

Before Nebius, you worked on products with a strong focus on ML. How has moving into infrastructure changed the way you think about building technology?

It’s completely different. Previously I was building user-facing B2C products, specifically search suggestions for one of the largest search engines in the world, with a heavy focus on ML, A/B testing, and a very short feedback loop between me and the user. I could use the product myself and immediately feel whether something was right.

Infrastructure is a different set of values entirely. What matters is predictability, reliability, and clear troubleshooting when things go wrong. It’s like electricity: you plug into the socket and you want it to work, 24/7, at a fixed frequency, with no surprises. Everyone already knows what “good” looks like. The hard part is achieving it. The infrastructure challenge is its own kind of interesting.

What have you had the opportunity to build or shape at Nebius?

My first major task was rebuilding the compute and network services team from almost nothing. We started with only one person in my team when Nebius was formed. So for the better part of the year, hiring was essentially my full-time job: hiring events, cold outreach, five to seven interviews a week, reaching out directly even without recruiters.

The turning point was when the company began establishing itself in Amsterdam in 2023. Amsterdam is the most international city in continental Europe, with great infrastructure from a financial and legal point of view.

What is Nebius like from the inside?

The team is strong. Seriously strong; the level is comparable to big tech in its best years. People work efficiently, and everyone is focused on what actually needs to get done.

On product vision, I’d describe it as deliberately absent, and I mean that in a good way. The strategy is: AI is coming, we know hardware capacity will be needed, but we don’t know exactly how the market will evolve, which services will dominate, or what the ratio of inference to training will look like in three years. We will do what’s needed. We are capable and efficient, and we move fast when the market shows us where it’s going.

Where we are now is above all possible expectations we had two years ago. The targets that seemed unreasonably ambitious back then, we are now well past them. That’s actually mind-boggling. That wouldn’t have been possible without the AI wave, but you still have to be built well enough to catch it.

How would you describe your work personality?

Results-oriented, delegating, and I’ll be honest, productively lazy. Being lazy in the right way means I try to be intentional about where I invest my time, while creating opportunities for others to take ownership, develop their skills and contribute meaningfully.

Any reflections on where the industry is heading?

I don’t think this is a bubble. If you’ve actually used these tools, asked ChatGPT about something you genuinely care about, used it as an advisor in an area where you’re not an expert, the value is real. I use it for medical questions. I recently used it to troubleshoot a plumbing problem, taking pictures of the pipes. It already works.

My bigger conviction is that the truly transformational moment will be physical AI, when models can act in the physical world. Humans, at the end of the day, want a fairly short list of things: enough food, entertainment, health, and to not spend their lives doing physical labor. Physical AI can address that last one. I believe physical AI will be a norm. If you take all cars, all home appliances, all smartphones and electronic devices combined, I think the physical AI market will eventually be larger than all of that. It won’t happen tomorrow, but I believe it’s coming. And we will have way more time to be humans, do human things: take care of our health, play sport, sing songs.