Departure index · 2019 = 100
outricher/Russian IT exodus
Analytical report · Russian-speaking professionals worldwide · REPORT 2026-RU-01 · EDITION 3 · DATA AS OF 12.06.2026
Diaspora/2026
Who left Russia, where they settled, and how the Russian-speaking professional world abroad works.
The first quantitative study built on 1.17 billion career profiles: the scale and dynamics of the exodus, a portrait of the diaspora by occupation and seniority, the inner workings of “Russian IT” abroad, an AI deep dive, and practical benchmarks for companies, HR teams and market players. Three independent signals, 30 countries measured, calibration against 69,142 known professionals, samples validated by exact counts.
1.2 M
professionals with Russian education abroad: 934K measured directly, 1,244K by model
200–600 K
left after February 24, 2022 and have not returned — a conservative model range
×2
2022 departures: double the 2021 pace and four times pre-war 2019 — the break shows in two independent signals
61%
of the war wave is IT and data: the most technical cohort in the country's history has left
8,182
people work for Russia's flagship employers while living abroad — “tethered”
Colour keythe diaspora as a wholethe 2022 waveplacespeople and roleseducation and skills
Summary / 12 findings
What we learned by counting the diaspora through its career footprints
- More than a million professionals with Russian education live and work outside Russia: 934,037 measured directly across 30 countries; the model puts it at 1,243,788. Every figure is a lower bound.
- Between 200,000 and 600,000 left after February 24, 2022 and are still abroad (187,505–615,856 by model). Consistent with The Bell (~650K emigrants of all kinds): our figure covers only people with a career profile.
- 2022 was the peak in both volume and non-return: departures doubled against 2021 (×2.0; ×4 against pre-war 2019), and the 2022 cohort shows the highest share still abroad — 64% of those who left the flagships in 2022 are abroad today.
- The war wave is the most technical: 61% of those who left the flagships in 2022+ are engineers and data specialists (against 28% across the diaspora). The industry's core left.
- The diaspora is a mature labor market, not a student body: 47% have 15+ years of experience; the median in Israel and the Netherlands is 16 years. New hubs run younger: Georgia's median is 10 years.
- Companies followed the people: 8,182 people with Russian backgrounds work for the flagships from abroad; Yandex accounts for 3,612, including 1,031 in Serbia.
- Serbia is the war wave's No. 1 destination (1,931 people from the flagship cohort), followed by Germany, Kazakhstan, the US and the UAE. Post-Soviet hubs absorbed the volume; the West took the seniors: 46% of the wave in Cyprus is senior+, against 32% in Kazakhstan.
- Cities: old emigration versus new. The largest communities are New York (≈32K) and the Bay Area (≈29K); but Tbilisi, Belgrade, Yerevan and Limassol grew almost from zero in two years — in Serbia, 60% of datable arrivals came after 2022.
- The AI diaspora is already visible in frontier labs: 3,198 profiles with Russian backgrounds across 13 leading AI companies — 189 at OpenAI, 99 at Anthropic. Nebius — a company that is 53% people with Russian pasts — is worth more than $50bn.
- HSE is emigration's top university, followed by MSU, MIPT, Bauman MSTU and ITMO. The route depends on the alma mater: 24% of MIPT graduates who left are in the US.
- The academic diaspora numbers in the tens of thousands: ≈17,256 scientists and lecturers with Russian education abroad (model); 7.0% of the diaspora works in universities and research.
- The diaspora is reachable: 75.7% of the cohort can be identified for direct contact, each with a full career history. This is the largest labeled pool of Russian-speaking senior talent outside Russia.
01
Scale
A diaspora the size of a metropolis
Emigration is usually estimated through surveys and border statistics. We followed career footprints instead: Russian education in the résumé, employment at flagship Russian companies, Russian language in the profile. Every figure is either a direct count across 1.17 billion profiles or a model with an open formula. The samples were validated by exact counts: they understate shares by 22–30%, so everything below is a lower bound.
EXHIBIT 01
MEASURED
The measurement ladder: from 6.3 million profiles in Russia to a million-strong diaspora
Three independent signals — three levels of confidence; we show all of them, not just the flattering one
| Rung | People | What it means |
|---|---|---|
| Profiles currently located in Russia | 6,305,071 | exact count — how many professionals are “visible” inside the country |
| Russian-educated abroad — measured | 934,037 | direct count across 30 countries: the profile lists a Russian university or school |
| Same, adjusted for profile completeness (model) | 1,243,788 | education is recognized for only 75.1% of even known-Russian professionals — we divide by that share |
| Listed Russian as a language — all measured countries | 2,125,034 | the broadest signal: includes Russian speakers from across the former USSR |
| — of them, outside the former USSR | 1,710,328 | Russian-speaking professionals beyond the post-Soviet space |
HOW TO READThe language signal is broader than the education signal but more sparsely filled in: only 39.5% of even elite Russian professionals list languages — so 2.1M “listed Russian” profiles are consistent with a diaspora of several million. Why it matters: every row can be re-verified by re-running the query against the data.
EXHIBIT 02
MEASURED
Geography: the US is the largest pool
Profiles explicitly listing Russian, by country — lower bounds
HOW TO READBlue bars — former USSR (historical communities + the new wave); black bars — the rest of the world.
EXHIBIT 03
MEASURED
Concentration: the most “Russian” markets
Share of a country's profiles listing Russian, %
HOW TO READIn Armenia, 12.9% of all profiles list Russian; in Georgia, 10.2% — against 0.2% in the US. This is a map of density, not size.
EXHIBIT 04
BENCHMARK
Our estimate against external ones: the numbers converge
These sources measured different things by different methods — which makes the agreement all the more valuable
| Source, year | What it measured | Estimate | Consistent? |
|---|---|---|---|
| Re:Russia, 2023 | everyone who left by mid-2023 | 820–920K | yes — our diaspora is the “has a profile” subset |
| The Bell, 2024 | those remaining abroad | ~650K | yes — our 200–600K covers professionals, without families |
| Russia's Ministry of Digital Development, 2022 | IT specialists who left | ~100K | yes — the IT share of the war wave is of the same order |
| Wachs, EPJ Data Science, 2023 | developer migration via code | 11.1% by Nov 2022 | yes — an independent digital footprint, same dynamics |
| OutRush, 2023–24 | emigration composition (surveys) | >45% in IT, 80% with degrees | yes — matches on structure |
| This study, 2026 | Russian-educated professionals abroad | 934,037 measured · 1.24M by model | direct measurement of career footprints, 30 countries |
WHY IT MATTERSNo single source sees the whole picture: surveys do not know the underlying population, and border statistics cannot tell a tourist from an emigrant. Career profiles are the only dataset where the person, their occupation and their new country are visible at once. We do not cite estimates with undisclosed methodology (chapter 07).
02
Dynamics
The 2022 wave: a break visible in any dataset
The year of departure is reconstructed from career history: when the last position with a Russian location ends, the person is “dated.” This yields two fully independent curves: one for the Russian-educated diaspora, one for exits from ten flagship employers. Both break in 2022.
EXHIBIT 05
MEASURED
Two independent signals, one break: departures double in 2022
Departure pace against the pre-war 2019 baseline (index, 2019 = 100): different scales — same shape
What the two signals are
Signal A — diaspora career histories. A person with Russian education closes their last Russia-located position — a dated "departure" event (5,665 datable events in the sample). 2022 has twice as many of them as 2021 — and four times the 2019 level.
Signal B — exits from the ten flagships. 69,142 people who left Yandex, Sber, T-Bank and the others since 2019. Leaving a company ≠ leaving the country, so the pace is lower — ×2.0 vs 2019; but a record 64% of the 2022 cohort ended up abroad (Exhibit 07).
Why the agreement can be trusted: the signals are built on different people, different data tables and different definitions — and both break at the same point, in early 2022. 49.5% of all datable departures in today's diaspora came after January 1, 2022.
HOW TO READThe index removes the scale gap (hundreds of sample events vs thousands of cohort people) and keeps the curve shapes. Hover for the underlying values. The 2023–2025 decline is partly real stabilization, partly lag: fresh departures enter career data late, so the right tail is understated.
EXHIBIT 06
MEASURED
The flagship exodus: 69,142 people since 2019
The ten largest tech-sector employers; “abroad” is measured among those whose current country is known
| Company | Left 2019+ | After Feb 2022 | Now abroad | Share abroad |
|---|---|---|---|---|
| Yandex | 23,686 | 17,813 | 62.4% | |
| Gazprom | 1,588 | 709 | 70.1% | |
| Kaspersky | 5,446 | 3,447 | 60.2% | |
| VK (Mail.ru) | 3,386 | 711 | 53.8% | |
| Sber | 12,500 | 7,578 | 53.3% | |
| Wildberries | 1,810 | 1,455 | 50.7% | |
| T-Bank (Tinkoff) | 9,195 | 7,470 | 49.9% | |
| Avito | 3,423 | 3,019 | 49.5% | |
| Ozon | 6,371 | 5,175 | 47.0% | |
| MTS | 6,751 | 3,860 | 39.3% |
HOW TO READThis is not ordinary churn: more than half of those who left are now abroad. One in six of the 69,142 came from Yandex. From the primary source: Yandex's founder told the FT that “half the top management and 10% of the developers” ended up outside Russia.
EXHIBIT 07
MEASURED
2022: the peak of volume and non-return
Exits from the flagships (bars) and the share who ended up abroad (line)
HOW TO READBefore the war, 55–59% of job-changers went abroad; in 2022 it was 64% at record volume: people quit in order to leave the country.
EXHIBIT 08
MEASURED
Where the war wave went
The 2022+ flagship cohort now abroad: countries
HOW TO READA two-layer map: visa-free hubs absorbed the mass quickly, while Western markets recruited through work visas — slower, but for good.
EXHIBIT 09
MEASURED
War-wave cities: Belgrade is the largest single hub
The 2022+ flagship cohort abroad, by city
CONTEXTSince February 2022, Russian citizens have registered 789 companies and 2,134 sole proprietorships in Serbia (official European Parliament reply, 2023). The IMF links Armenia's 12.6% GDP growth in 2022 to the inflow of people and capital from Russia; remittances to Georgia grew fivefold, to $2bn a year.
03
Portrait
Who these people are: occupations, seniority, universities
We classified 12,916 Russian-educated profiles across 30 countries: job titles, seniority, universities, fields of study, cities. The summary distributions are country-weighted — without this, fixed samples would overstate small hubs and understate the US and Europe. The diaspora is not “a hundred thousand IT workers”: it is an economy in miniature, with doctors, lawyers, scientists and entrepreneurs.
EXHIBIT 10
MEASURED
What the diaspora does: IT and data at 28%, the rest spans the whole economy
Distribution by function among profiles with a recognized role; country-weighted
Major blocks · % of recognized roles
Full distribution · % of recognized roles
HOW TO READIT engineering plus data/AI is the largest block but not a majority: next come executives and business owners, marketing and sales, operations and finance. Why it matters: for host countries this is an inflow not only of developers but of managers, doctors and teachers.
EXHIBIT 11
MEASURED
Seniority: the seniors left, not the students
Diaspora seniority, % of recognized; country-weighted
HOW TO READIndividual contributors are the majority, but one in six is a head-level-or-above manager or a founder.
EXHIBIT 12
MEASURED
Where the seniors go: war-wave seniority
Share of senior+ among 2022+ flagship leavers, by country
HOW TO READCyprus, Portugal and the UAE collect the top of the pyramid; Kazakhstan and Belarus take a younger wave. The gap is nearly twofold.
EXHIBIT 13
MEASURED
The diaspora's alma maters: HSE has overtaken MSU
Left — universities by graduate count in the sample; right — what they studied (field-of-study mentions)
Universities · graduates in the sample
Fields of study · mentions
4%
of profiles listing a degree include a PhD, a Candidate or a Doctor of Sciences
1.4%
list an MBA — management education on top of a first degree
74%
graduated in the 2010s or later — the diaspora's core is aged 28–38
CONTEXTMacroPolo's Global AI Talent Tracker ties Russia's growing share of the world's top-tier AI researchers to a single university — MIPT. The fields of study agree: computer science, applied mathematics and economics top the list.
EXHIBIT 14
MEASURED
The university sets the route: MIPT graduates head for the US, HSE graduates go everywhere
Where top universities' graduates end up: country shares among a university's graduates found abroad (weighted), %
| University | In sample | Top destinations (share of the university's graduates, %) |
|---|---|---|
| HSE University (Higher School of Economics) | 731 | United States 13.4% · Germany 13.4% · UAE 9.2% · United Kingdom 7.8% · Spain 7.3% |
| Lomonosov Moscow State University | 640 | United States 32.3% · Germany 10.3% · Kazakhstan 6.2% · Italy 5.9% · Netherlands 5.8% |
| Saint Petersburg State University | 422 | United States 19.3% · Germany 17.5% · Netherlands 6.3% · Canada 5.4% · Spain 5.3% |
| Bauman Moscow State Technical University | 286 | Germany 15.7% · UAE 13% · Canada 10% · Georgia 7% · Israel 5.7% |
| Peter the Great St. Petersburg Polytechnic University | 279 | United States 14.1% · Türkiye 8.4% · Germany 8.2% · Finland 8.1% · Spain 7.7% |
| ITMO University | 255 | United States 27.4% · Germany 13.2% · Canada 7.6% · Georgia 5.5% · United Kingdom 4.8% |
| Plekhanov Russian University of Economics | 248 | Germany 13.1% · Netherlands 11.8% · Canada 9.4% · UAE 7% · Armenia 6.1% |
| RANEPA (Presidential Academy) | 178 | UAE 17% · Italy 15.2% · Spain 12.8% · Kazakhstan 10.4% · France 7.6% |
HOW TO READPercentages are within each university's row. Technical universities gravitate toward the US and Northern Europe; economics-focused ones toward the UAE and Cyprus. A university's “career gravity” is the most stable predictor of destination country — stronger than occupation.
EXHIBIT 15
MODEL
Diaspora cities: New York and the Bay Area ahead of Tbilisi and Belgrade
Estimated Russian-educated diaspora by city, people (weighted model)
HOW TO READOld emigration built the largest communities in New York and California; the war wave built new hubs — Tbilisi, Belgrade, Yerevan, Limassol — almost from scratch in two years. In Germany the diaspora is scattered across five cities; in Serbia it is concentrated in one. For anyone opening offices or assembling teams, this is a direct pointer.
04
Current state · for companies and HR
How the diaspora has settled: employers, experience, market maturity
The exodus is half the story. The other half is how these people live and work now: who employs them, how much experience they carry, which communities have settled and which are still forming, and who keeps working for Russian companies from abroad. This is an operating map for anyone hiring, opening offices or building products for the diaspora.
EXHIBIT 16
MEASURED
Who the diaspora works for
Current employer categories, % of the diaspora; country-weighted
HOW TO READMost are embedded in ordinary local and international companies in their host countries; 7.0% work in universities and research, 2.1% in global corporations from the brand list, 1.8% freelance. Categories follow brand dictionaries, conservatively.
EXHIBIT 17
MEASURED
Experience: a market of seasoned professionals
Years since first career position, % of the diaspora; weighted
HOW TO READ47% of the diaspora has 15+ years of experience. Juniors are scarce: the young either did not leave or are not yet visible in career data.
EXHIBIT 18
MEASURED
Market maturity: median experience by country
Median career years of the diaspora in each country
HOW TO READIsrael, the Netherlands, Switzerland and Finland are “old” markets with a median around 16 years. Georgia and Armenia hold the young wave: a median of 10–10 years. Experience is a proxy for age and salary expectations.
EXHIBIT 19
MEASURED
Community age: where the diaspora is “new”
Share of datable arrivals after January 1, 2022, % by country
HOW TO READIn Serbia and the UAE, ~60% of the community arrived after 2022 — the infrastructure (schools, services, communities) is still forming. In Finland and the US, old emigration dominates (26% newcomers in Finland).
EXHIBIT 20
MEASURED
“Tethered”: 8,182 people work for Russia's flagship employers from abroad
Current employees of the 10 flagships with Russian backgrounds outside Russia; local hires at overseas offices filtered out
By company · people with RU backgrounds
What it means
Yandex leads: 3,612 people, 1,031 of them in Serbia — the company's second engineering center is now effectively in Belgrade. At T-Bank and Avito the count runs into the hundreds.
Why it matters: remote work for Russia is the forgotten part of the exodus. These people pay taxes in their new countries while working for Russian businesses. For local markets they are a talent reserve with a known employer; for Russian companies, a live network abroad.
The count is exact: all 22,496 foreign-based profiles of current employees were checked; a Russian background was confirmed for 36%.
HOW TO READThe top geography of Yandex's “tethered” staff — Serbia, Armenia, Kazakhstan, the UK — is, in effect, the map of its hubs abroad.
EXHIBIT 21
MEASURED
The diaspora's employers and industries
Left — the largest current employers in the sample; right — industries of employment
Employers · people in the sample
Industries · people in the sample
HOW TO READThree layers of employers: international outsourcers and fintech (EPAM, Exness, Revolut), relocated RU-rooted companies (inDrive, JetBrains, Wargaming, Playrix), and a visible freelance segment. Why it matters: the diaspora is already embedded in global companies — a channel for transferring expertise that sanctions cannot close.
05
The headline segment
IT under the microscope: who writes the code, and where
IT is the core of the exodus and the market's main interest: 61% of the flagship war wave are engineers and data specialists. We break this mass down by occupation, country, skills and new employers.
EXHIBIT 22
MEASURED
The diaspora's IT occupations
% of the IT cohort; country-weighted
HOW TO READBackend and analytics are the bulk; ML/Data Science is the growing core.
EXHIBIT 23
MEASURED
War-wave IT occupations
Flagship leavers 2022+, people
HOW TO READThe flagships were losing data and platform teams — the wave's mix is visibly more data-science-heavy than the older diaspora's.
EXHIBIT 24
MEASURED
Engineering hubs versus service hubs
IT and data share of each country's diaspora, %
HOW TO READGeorgia, Armenia, Serbia, Cyprus and the Netherlands are “engineering” hubs: people moved there in whole teams. In Turkey and Spain the diaspora leans more toward services and entrepreneurship.
EXHIBIT 25
MEASURED
Each wave more technical than the last
Share of IT/data and AI roles among flagship leavers, by year
HOW TO READThe “quality” of the outflow grew faster than its volume: technical concentration peaked in 2022–2023, and AI roles are rising in both share and absolute numbers.
EXHIBIT 26
MEASURED
Where the IT wave landed and what it can do
Left — employers of 2022+ leavers (intra-group transfers excluded); right — skills of the diaspora's IT cohort
The war wave's new employers · people
IT diaspora skills · mentions in the sample
Routes out of Yandex · war wave, top destinations
| Destination | People | Destination | People |
|---|---|---|---|
| Yandex (group) | 769 | Nebius | 210 |
| Avride | 128 | Yandex Practicum (group) | 115 |
| 107 | Yango | 99 | |
| JetBrains | 97 | Yandex Practicum (group) | 97 |
HOW TO READThree route types: relocated RU-rooted companies (Nebius, inDrive, JetBrains), global tech (Google, Microsoft, Booking, Revolut) and intra-group transfers — employees who moved with their companies' international entities. Skills are self-reported: SQL, JavaScript, Java, Linux, C++ — the portrait of a solid production engineer.
06
The most expensive talent
The AI deep dive: the Russian-speaking footprint in the model race
We checked every profile listed at the world's leading AI companies for a Russian background — education in Russia or Russian language in the résumé. This is direct measurement, not extrapolation.
EXHIBIT 27
MEASURED
Frontier companies: 3,198 people with Russian backgrounds
Every profile listing the company as current employer checked one by one; “RU signal” = Russian education or Russian language
| Company | Total profiles | RU signal | Share | Where they sit (RU cohort's top countries) |
|---|---|---|---|---|
| NVIDIA | 57,951 | 1,593 | 2.7% | United States, Israel, Germany |
| Nebius | 1,288 | 685 | 53.2% | Netherlands, Israel, Germany |
| Google DeepMind | 64,539 | 362 | 0.6% | United States, Canada, Mexico |
| OpenAI | 10,983 | 189 | 1.7% | United States, United Kingdom, Germany |
| Anthropic | 4,805 | 99 | 2.1% | United States, United Kingdom, Ireland |
| Perplexity | 1,380 | 96 | 7.0% | Serbia, United States, United Kingdom |
| xAI | 3,572 | 90 | 2.5% | United States, United Kingdom, Germany |
| ElevenLabs | 779 | 32 | 4.1% | United Kingdom, Germany, United States |
| Cohere | 1,151 | 24 | 2.1% | Canada, United Kingdom, United States |
| Hugging Face | 7,517 | 9 | 0.1% | Russia, India, United States |
| Stability AI | 252 | 8 | 3.2% | United States, Germany, Ukraine |
| Midjourney | 186 | 7 | 3.8% | Türkiye, United States, Russia |
| Runway | 362 | 4 | 1.1% | Russia, United Kingdom, Indonesia |
HOW TO READ“Total profiles” means people listing the company as their current employer; for high-profile brands (DeepMind, Hugging Face) this noticeably exceeds actual headcount — judge by the RU cohort's absolute numbers, not the shares. The signal is conservative: without a Russian university or the Russian language, a person is not counted — the real numbers are higher. Nebius is a special case: the company, which grew out of Yandex's international assets, is half made up of people with Russian pasts and is worth more than $50bn (mid-2026).
EXHIBIT 28
MODEL
Big Tech: ≈21.5K Russian speakers across the five largest companies
Exact count of employee profiles + RU share from a 30K-per-company sample → estimate
RU cohort estimate · people (model)
Calculation
| Company | Profiles | RU share | Estimate |
|---|---|---|---|
| 487,580 | 1.5% | ≈7,476 | |
| Microsoft | 664,654 | 1.0% | ≈6,802 |
| Apple | 466,726 | 0.7% | ≈3,158 |
| Amazon | 209,718 | 1.6% | ≈3,369 |
| Meta | 84,569 | 0.9% | ≈755 |
Samples of this type understate the share by 22–30% (validated by exact counts) — the estimates are conservative.
WHY IT MATTERSRussian-speaking engineers are a stable part of Big Tech's core workforce, comparable to the headcount of an entire IT company.
EXHIBIT 29
MEASURED
AI roles in the diaspora: ≈18,185 specialists — and the share grows with every wave
ML engineers, data scientists and researchers in the diaspora and the war wave
≈18,185
AI specialists in the diaspora by the weighted model — 1.9% of all who left
9%
of the flagship war wave holds AI roles: in 2022, whole data teams left at once
~400
former Yandex AI engineers are building Nebius — per the FT; our profile count confirms it
The Russian AI school is, above all, MIPT, HSE and the Yandex School of Data Analysis. After 2022 that school physically moved: its alumni now publish from London, Amsterdam and San Francisco. Our count across Nebius profiles: 685 of 1,288 employees (53%) have Russian backgrounds — the largest “Russian-speaking” asset in the global AI economy since the split of Yandex (a $5.2bn deal, July 2024).
WHY IT MATTERSAI talent is the defining shortage of the decade. The Russian-educated diaspora is one of the largest non-American pools of that talent, already redistributed across Western and Middle Eastern jurisdictions.
07
Practice + methodology
Benchmarks for the market — and how it was all counted
The diaspora is a market: the largest labeled pool of Russian-speaking senior talent outside Russia. This chapter holds the summary tables for decision-making and the full methodology with its limitations.
EXHIBIT 30
MEASURED
Country profiles of the diaspora: a summary table for decisions
Countries with a sufficient sample: structure, seniority, entrepreneurship, top city
| Country | Diaspora (floor) | IT share | Senior+ | Founders | Top city |
|---|---|---|---|---|---|
| United States | 444,023 | 9% | 20% | 2.4% | New York |
| Germany | 217,400 | 22% | 21% | 4.2% | Berlin |
| Poland | 109,569 | 20% | 27% | 3.8% | Warsaw |
| United Kingdom | 99,651 | 19% | 28% | 4.3% | London |
| Türkiye | 70,050 | 13% | 18% | 6.5% | Istanbul |
| Spain | 64,660 | 20% | 22% | 4.2% | Barcelona |
| Israel | 63,561 | 30% | 21% | 3.9% | Tel Aviv |
| Georgia | 59,070 | 35% | 27% | 3.8% | Tbilisi |
| Kazakhstan | 58,167 | 20% | 28% | 2.0% | Almaty |
| Netherlands | 55,001 | 14% | 15% | 4.9% | Amsterdam |
| Armenia | 46,307 | 30% | 28% | 4.2% | Yerevan |
| Switzerland | 44,715 | 13% | 23% | 7.5% | Zurich |
| UAE | 44,027 | 15% | 32% | 7.5% | Dubai |
| Czechia | 41,225 | 21% | 30% | 1.8% | Prague |
| Finland | 39,171 | 17% | 17% | 3.0% | Helsinki |
HOW TO READ“Diaspora (floor)” is the lower bound: profiles explicitly listing Russian. IT share, senior+ and founders come from the classified sample within each country. The key arbitrage: comparable senior talent costs differently in Belgrade and London — the table shows where concentration runs highest.
EXHIBIT 31
MEASURED
Where to find whom: a concentration map by occupation
Top countries by share of each function's holders (weighted)
| Who you are looking for | Where concentration is highest (share of the function's holders, %) |
|---|---|
| Data & AI (≈25,305 (model)) | United States 18.8% · UAE 7.1% · Netherlands 7% · France 6% · Germany 5.5% |
| IT engineering (≈117,362 (model)) | United States 27% · Germany 10.2% · Serbia 5.3% · Georgia 5.2% · Kazakhstan 5.2% |
| Design (≈21,680 (model)) | United States 14.6% · Germany 11.3% · Georgia 6.2% · Spain 6% · Netherlands 5.1% |
| Product (≈11,978 (model)) | United States 13.2% · Germany 12.8% · United Kingdom 9.3% · Netherlands 9.2% · Spain 7.2% |
| Marketing & PR (≈35,449 (model)) | United States 31.3% · Germany 9.5% · Indonesia 5.6% · UAE 5.5% · Spain 4.9% |
| Finance (≈30,240 (model)) | United States 52.4% · UAE 4.9% · Kazakhstan 3.8% · Netherlands 3.7% · Canada 2.9% |
| Science & academia (≈13,682 (model)) | United States 58% · Germany 9% · Israel 3.3% · Netherlands 3.2% · Spain 3.2% |
| Media & content (≈22,264 (model)) | United States 49.9% · Türkiye 5.7% · Italy 4.6% · Indonesia 4.5% · Canada 3.9% |
| Executives & owners (≈30,530 (model)) | United States 36.4% · Kazakhstan 7.8% · UAE 7.5% · Germany 4.5% · Spain 4.2% |
| Medicine (≈10,028 (model)) | United States 47.4% · Canada 8.8% · UAE 8.1% · United Kingdom 5.5% · Italy 5.1% |
PRACTICAL TAKEAWAYAssemble an ML team in the US, London and Belgrade; game development in Cyprus and Serbia; marketing and content in Turkey, the UAE and Spain; research in the university centers of the US and Germany.
EXHIBIT 32
MEASURED
How reachable the diaspora is: contact coverage
Measured on a cohort of 16,677 diaspora profiles joined to the contact data layers
75.7%
of the cohort can be identified for direct contact through the business-email layer — measured, not estimated
11.7%
carry an extended dossier: personal addresses, social accounts, employment beyond the résumé
100%
are covered by a full career profile: dated position history, education, skills, languages — 95+ fields
What the data lacks and what we refuse to do: phone numbers are practically absent for this cohort (<1%), and we do not “reconstruct” them; missing values ship as NULL, not as a guess.
WHY IT MATTERSThe usual line about the diaspora is that “these people can be neither seen nor surveyed.” This population is visible: measurable, segmentable and reachable with known coverage.
Subject
Russian-speaking professionals who left Russia; working definition — career footprints: Russian education, employment at flagship companies, Russian language in the profile
Dataset
1.17 billion career profiles; measurements across 30 countries; 12,916 diaspora profiles classified in detail
Signal 1 · Education
Text classification of educational institutions (Cyrillic + transliteration, 52 named universities + generic patterns). University identifiers are sparse in the data — classification is text-only
Signal 2 · Employer
Dated positions at 10 flagships (Yandex, Sber, T-Bank, MTS, Kaspersky, Ozon, Avito, VK, Gazprom, Wildberries); 69,142 unique leavers since 2019
Signal 3 · Language
Russian in the languages block (all spelling variants). Weak but massive: even among known-Russian professionals, only 39.5% list languages
Calibration
Against 69,142 known-Russian professionals: education is filled in for 93.1% and recognized as Russian for 75.1% — the coefficient for converting measured floors into model estimates
Validation
Sample shares were validated by exact counts in three countries: samples understate by 22–30% — all sample figures are published as lower bounds, uncorrected
Weighting
Summary distributions are country-weighted: weight = (the country's Russian-education share × profile population) / sample rows — otherwise fixed samples overstate small hubs
Departure dating
The year the last position with an explicit Russian location ended; 5,665 datable departures; 49.5% after January 1, 2022
Edition
3rd: adds the “How they have settled” chapter, experience, employer categories, community age, skills. All data queries ran on June 12, 2026
Why the figures are lower bounds. Every step is conservative: samples underrepresent new profiles (we apply no correction); the education signal loses a quarter of even known-Russian résumés; the language signal catches only one in three; a person with no recognized signal is not counted at all. The diaspora's true size exceeds the published estimates — a deliberate choice in favor of verifiability.
What we do not claim. We do not measure citizenship or ethnicity — only career footprints. We cannot tell “left for good” from “has long lived abroad” outside datable cases. We do not add overlapping signals into a single “grand total.”
Platform coverage. Career profiles see internationally oriented white-collar workers; blue-collar trades and offline businesses are underrepresented.
Update lag. Profiles get updated late — the 2024–2025 tail of the departure curves is understated more than other points.
Location is self-reported. Digital nomads and dual bases add noise in both directions.
The post-Soviet commonality. The language signal includes Russian speakers without Russian backgrounds; for strict estimates, use the education signal.
Profiles ≠ headcount. “Company employees” are people listing it as their current employer; close to headcount, but not equal to it.
Classifiers. Occupations and seniority are recognized from job-title text; rare phrasings fall into “other.”
External sources (vetted for citability):
Re:Russia (2023) — 820–920K who left by mid-2023 · The Bell (2024) — ~650K remaining abroad · Russia's Ministry of Digital Development (2022) — ~100K IT specialists who left · J. Wachs, EPJ Data Science (2023) — 11.1% of Russian open-source developers changed country by November 2022 · OutRush / EUI (2023–24) — ≤8% returned; >45% of the emigration in IT; 80% with degrees · Novaya Gazeta Europe (2024) — ≥2,500 scientists severed Russian affiliations · FT (2024) — the $5.2bn split of Yandex's assets; ~1,300 employees in the international part · MacroPolo Global AI Talent Tracker 2.0 (2024) — Russia's growing share of top-tier AI researchers, the role of MIPT · IMF (2023) — Armenia's GDP +12.6% in 2022 on the inflow from Russia · European Parliament (2023) — 789 companies and 2,134 sole proprietorships registered by Russian citizens in Serbia.
What we deliberately do not cite: “4 million who left” (border crossings, not people), extrapolations from small surveys, “two-thirds of IT specialists have returned” (methodology undisclosed; independent surveys put it at ≤8%), and SIM-card counts (which cannot tell a return from a visit).
Outricher Research
Need this cut for your market?
Every table and chart here can be recomputed on request, and cut further — by country, industry, occupation, employer or cohort. Tell us which question you need answered; we reply within one working day.