WHERE YANDEX'S
TECHNICAL TALENT
WENT.
One of the world's best engineering organisations came under pressure. 11,828 of its engineers moved. They didn't scatter: they clustered. And almost nobody is competing for them.
A talent pool that clusters instead of scattering turns a hard search into a question of going to the right place. This report shows you where that place is, and what to do when you get there.
THE MOVEMENT TIMELINE
Year is when each person started their current role. The shape is clear: a step-change in 2022, and a steady climb every year since. This is a live market, not a historical event.
*2026 is a part-year. New roles started each year, based on 10,608 people with a datable current-role start.
SO WHAT?
The flow hasn't stopped: it's accelerating. 2,964 new roles started in 2025, up from 567 in 2021. If your hiring plan treats this as a one-off pool, you're already behind the companies treating it as a live market.
WHERE THEY LANDED
Most of the volume is in the long tail: scale-ups, dev-tools and relocation employers. Only 155 landed at frontier AI labs. Which means the vast majority of this pool is not where everyone is already looking.
Scale-ups, dev-tools, fintech & the long tail
Big tech: Google, Meta, Microsoft, Amazon, Apple, Booking
AI infrastructure & cloud: Nebius, Wayve
Frontier AI labs: Perplexity, DeepMind & peers
SO WHAT?
Only 155 of 11,828 are at frontier AI labs, less than 2%. The other 98% are reachable, placed at employers who aren't your direct competitors for them. The question isn't whether this pool is accessible. It's whether you get there before the labs that know it exists.
THE GEOGRAPHIC MAP
The diaspora pooled by city, and the cities split into two types: established tech capitals you already know, and relocation hubs that filled up fast from 2022 onward. Most sit in European time zones.
SO WHAT?
Belgrade, Yerevan, Tbilisi and Almaty hold over 1,100 engineers between them, in European time zones, with strong technical depth, and almost zero recruiting competition. London is crowded and expensive. These four cities are where the same quality talent waits, unchallenged.
WHO THEY ARE
Across 3,800 profiled in detail, the distribution weights heavily toward experienced individual contributors and senior engineers, with a real staff-and-above layer on top.
SO WHAT?
This isn't a graduate pool you develop over 18 months. These are people with proven track records at one of the world's best engineering organisations. The hiring question isn't "can they do the job": it's "can you get in front of them before someone else does."
THE FOUNDER FACTORY
This group does not only join other people's companies. 561 now run their own, with 22 building in stealth right now.
now run their own company. The most entrepreneurial members of this pool have already left, to build.
are building in stealth right now. These companies recruit from the same pool you are.
SO WHAT?
The window is closing. The most entrepreneurial engineers have already moved. The next wave, the ones still in senior IC and staff roles at scale-ups, need to be reached now, before the alumni-founded companies find them first.
THE QUANT PULL
152 of this group went to quant and trading firms. You already have competition you may not have accounted for.
SO WHAT?
Quant firms got here early. If your role can't match finance money, the only thing that wins is a harder, more interesting problem, and the equity to match it. Know this before you open the search.
RESEARCHER SPOTLIGHT
Within the 11,828, there is a tighter cohort: 263 people whose work is genuinely ML research, the cohort an AI lab actually competes for.
Genuine ML researchers or applied AI specialists. The subset an AI company is actually competing for.
SO WHAT?
If your next hire is a research-grade ML engineer, London is where you look, specifically inside DeepMind, Meta Research and Microsoft Research. The 263 are mapped. They are approachable. The question is whether you have the access and the verified signal to get to the right 20 before a competing lab does.
WHAT THIS MEANS FOR YOUR NEXT HIRE
A map is only useful if it tells you what to do. Five reads, for anyone hiring from this pool right now.
The senior talent is already placed. Think lateral.
83% are senior or experienced individual contributors. This is not a graduate pipeline. These are people who can ship from week one.
The relocation hubs are the under-contacted opening.
Belgrade, Yerevan, Tbilisi and Almaty hold thousands of strong engineers who relocated under pressure and sit in European time zones. Almost nobody is recruiting into them.
You are competing with the founders.
561 of this group now run companies. The most entrepreneurial left to build, and their startups recruit from the same pool you do.
The quant desks got there first.
152 went to quant and trading firms, led by Pinely. For an AI lab, your competition for the strongest algorithmic engineers includes finance money.
The research-grade cohort is in London, at the labs.
The 263 genuine ML researchers concentrate in London, inside DeepMind, Meta and Microsoft Research. If you want research-grade talent from this background, London is the search, not a fallback.
WE'VE MAPPED
ALL 11,828.
WHO GETS
THERE FIRST?
We know which of these engineers are research-grade. We know which cities are under-recruited. We know which ones are open to a move right now.
Tell us the role: the level, the specialisation, the location constraint. We'll show you who's in your hiring zone from this pool: verified profiles, scored by fit, ready to approach.