AI/ML engineers with foreign education — the employment-visa demographic, computed.

No recruiting tool has a filter for "degree earned abroad" — so we compute it. Every education record in the professional-profile graph carries an institution; institutions resolve to countries; a US-based machine-learning engineer whose degree comes from outside the US belongs to the population that lives on employment visas or needs the next one — H-1B pressure today, an O-1 upgrade tomorrow, EB-1A / NIW as seniority grows. This cohort computes that population directly from education and career records and ships it ranked, with verified contacts. It is a recurring series: re-ordered five times by the same buyer.

5
batches re-ordered by one client
13,188
rows shipped in the series
44,581
qualified in the latest pool
100%
email coverage latest batch

recurring series · latest batch 2026-07 · MEASURED · Outricher data

Why they're likely clients. A US-based ML engineer whose degree was earned abroad belongs, by career structure, to the population that lives on employment visas or needs the next one — H-1B pressure today, an O-1 or EB-1A conversation as seniority grows. Every row ships as a ranked propensity signal with the education evidence attached, for an attorney to qualify — never an eligibility verdict.
Who buys this · and what it replaces

Employment-visa practices whose caseload runs on H-1B engineers; O-1 practices upgrading H-1B seniors before the lottery does it for them; and immigration-savvy recruiting firms and staffing agencies that sell visa-friendly pipelines to employers. What it replaces: guessing visa dependence from surnames and hoping — this cohort derives it from education records, evidence attached to every row.

The signal logic · three computed filters
S1 · structural

Degree geography

Every education record carries the institution; institutions resolve to countries. A US-based ML engineer whose bachelor's is from abroad is the employment-visa demographic by construction — no self-reporting, no keyword guessing.

→ the population no filter can reach
S2 · structural

Function precision

Machine-learning roles detected from titles and career history, not self-tagged skills. The latest funnel: 1M US-based engineering profiles scanned → 312,990 in ML-relevant roles → 44,581 with verified foreign education.

→ 1M → 312,990 → 44,581
S3 · derived

Seniority timing

Years of experience computed from dated positions — the graph holds 5.6 billion of them. The 8–15-year band is where O-1 / EB-1A conversations start replacing H-1B renewals; the cohort ranks toward it.

→ O-1 · EB-1A · NIW window
institution geography per education record degree level ML-role detection from titles & career history years of experience from dated positions US-based residence, person-level self-tagged skills — never used as a filter
Sample rows · drawn from the latest shipped batch

Twenty real rows from the latest 2,000-lead batch. Big-employer names are shown as-is; personal details are masked on this public page — the full rows, with names and verified emails, ship in the dataset (and a free 10-profile sample is one email away).

PersonRoleCompanyBased inEducationExperienceEmail
Mert S.AI Machine Learning Engineer StaffLockheed MartinAustin, Texas, United StatesBachelor's Degree — Ahmet Yesevi Üniversitesi17 yrsm***@live.com
Sam R.Staff Machine Learning Engineer/Data ScientistDatabricksCambridge, Massachusetts, United StatesScotch College → Doctor of Philosophy (Ph.D.), Massachusetts Institute of Technology6 yrss***@gmail.com
Charlelie L.Senior Scientific Machine Learning EngineerNVIDIASanta Clara, California, United StatesENS Paris-Saclay → Doctor of Philosophy - PhD, Toulouse INP10 yrsc***@***.com
Kaiyan Z.Senior Machine Learning EngineerSalesforceSan Francisco Bay AreaThe University of Hong Kong → Master's degree, Columbia University in the City of New York6 yrsk***@***.com
Mridul K.Senior ML EngineerUberSan Francisco Bay AreaMaster of Technology (M.Tech.) — Indian Institute of Technology, Roorkee8 yrsm***@***.com
Zhongmou L.Staff Machine Learning EngineerReddit, Inc.Greater Seattle AreaRutgers, The State University of New Jersey-Newark4 yrsl***@yahoo.com
Carlos G.Staff Machine Learning EngineerPrivate company · ~109 empUnited StatesInstituto Tecnológico y de Estudios Superiores de Monterrey → École Polytechnique Fédérale de Lausanne4 yrsj***@gmail.com
Nikhil W.Member of Technical Staff, Machine Learning Engineer IIVMwarePalo Alto, California, United StatesUniversity of Mumbai → University of Southern California4 yrsn***@***.com
Simon W.Senior ML EngineerAppleUnited StatesTechnical University of Berlin → Creative Destruction Lab11 yrss***@***.art
Siva M.Senior Machine Learning EngineerZooxSan Francisco Bay AreaInternational Institute of Information Technology → Doctor of Philosophy - PhD, Heidelberg University12 yrsk***@gmail.com
Arjun A.Senior Machine Learning EngineerAppleSanta Clara, California, United StatesNational Institute of Technology Calicut → The University of Texas at Austin4 yrsa***@***.edu
Gaurav S.Senior AI EngineerThe State University of New YorkSan Francisco Bay AreaInternshala Trainings → Master of Science - MS, University at Buffalo5 yrsk***@gmail.com
Sharhad B.Senior AI and ML EngineerAppleNew York, New York, United StatesFourthBrain14 yrss***@***.com
Patrick H.Senior Autopilot Machine Learning EngineerTeslaPalo Alto, California, United StatesKarlsruhe Institute of Technology (KIT) → Doctor of Philosophy - PhD, Technical University of Munich10 yrsh***@***.org
Ruixuan(Corey) D.Senior Machine Learning EngineerGoogleSan Francisco Bay AreaTsinghua University → Ph.D., Washington University in St. Louis4 yrsd***@***.com
Manoj A.Senior Data Scientist & Machine Learning EngineerIntel CorporationAustin, Texas, United StatesInstitute Of Chemical Technology → Executive MBA, Quantic School of Business and Technology5 yrsm***@gmail.com
Varsha S.Senior Machine Learning EngineerAdobePalo Alto, California, United StatesSri Sivasubramaniya Nadar College Of Engineering → Master of Science (MS), Stanford University4 yrsv***@***.com
Anuprit K.Senior Manager, Machine Learning EngineeringSalesforceNew York, New York, United StatesCollege of Engineering Pune5 yrsa***@gmail.com
Naveen K.Senior Machine Learning EngineerPrivate company · ~6 empNashville Metropolitan AreaVellore Institute of Technology → Master of Science - MS(Thesis), University of Georgia - Franklin College of 5 yrsn***@***.com
Adam P.Senior ML EngineerAppleSan Francisco Bay AreaGordonstoun → Master's, Machine Learning (Computer Science), UCL4 yrsp***@***.com

Rows are real, drawn from the latest shipped batch. Large-employer names are shown (a masked engineer at a large company is not identifiable); smaller employers are masked to industry and size. The Education column shows the foreign degree first — where an arrow follows, it traces the classic arc: foreign undergraduate → US graduate school. Masking: surnames reduced to initials, email domains hidden except free providers. Nothing here is generated or blended.

Why this works · the recurring proof

This exact cohort has been re-ordered five times by the same buyer — the strongest quality signal a dataset can have. Every batch ships deduplicated against everything previously delivered: a dedup ledger travels with the client, so a re-order is always fresh rows, never recycled ones. The latest batch: 2,000 leads at 100% email coverage and 129 avg fields per lead, drawn from the 44,581-strong qualified pool.

The honesty boundary

Degree geography is a proxy for visa dependence, not a visa-status record — current status is not observable in any dataset, including ours. The cohort is a ranked propensity pipeline, evidence-tagged, for your intake to qualify: signals, not adjudicated outcomes.

The series in numbers · measured across five shipped batches

Cross-batch totals are measured on the delivery files themselves: 13,188 rows shipped, 11,761 unique people (every batch deduplicated internally; the ledger keeps re-orders fresh). Deep analytics below are computed on the latest batch of 2,000 — the denominator for every percentage.

The shipped series · rows per batch
7,188wide batch
2,000latest batch (2026-07)
2,000earlier batch
1,000config batch
1,000refinement batch
13,188 rows · 11,761 unique people · latest batch 96.6% fresh

Reading: a client who re-orders five times is running this as infrastructure, not an experiment. The latest batch is 96.6% fresh versus everything before it — the dedup ledger works.

Origin universities · latest batch
University of Mumbai56
Waterloo40
BITS Pilani31
VIT30
Tsinghua28
JNTU25
Manipal23
Peking University17
University of Toronto17
US destination schools · graduate degrees
Columbia87
Carnegie Mellon67
Georgia Tech61
Northeastern61
USC39
Stanford30
MIT21
Master's · 1,325 of the batch (66.3%)
PhD · 263 of the batch (13.2%)

Degree ceiling on the batch: PhD 263 (13.2%), master's 1,325 (66.3%) — roughly four in five hold a graduate degree, the arc adjudicators recognize instantly.

Current employers · latest batch
Apple107
Google64
Meta46
Amazon45
TikTok41
NVIDIA40
Tesla26

Big-tech share 31.1%; half the batch sits inside 20,000+-employee companies. These are employed, senior, paid specialists — not job seekers.

Role families · latest batch
ML Engineer754
Data Scientist691
AI Engineer237
Data Engineer198
DL / CV / NLP specialist77

46.3% carry a senior / staff / principal / lead modifier. Geography: California-attributable ≥36.7%, New York ≥11.6%, Washington ≥8.7%, Texas ≥7.8%.

Evidence & momentum · latest batch
certifications on record994
honors & awards639
publications551
patents104

Evidence extracted from the 1,560 enrichment-matched rows (78%) — shares of the full 2,000 are lower bounds: publications 27.6%, awards 32.0%, patents 5.2%. Momentum: 43.7% started their current role within 18 months, median 8 years of experience — exactly the window where a status conversation lands. Email 100% (delivery gate), phone 6%.

Composite portrait
The foreign-educated ML engineer
Assembled from the medians and modes of the latest 2,000-lead batch.

The arc

  • Foreign first degree → US graduate school
  • Destination schools led by Columbia · Carnegie Mellon · Georgia Tech

Now

  • ML engineer / data scientist role
  • Employers incl. Apple 107 · Google 64 · Meta 46

Momentum

  • 43.7% of the batch changed roles within the last 18 months
  • Certifications 994 · publications 551 on the enrichment-matched base
13,188 rows shipped11,761 uniquerecurring series
Composite portrait — assembled from cohort medians and modes; not a real individual.
Get this dataset

Everything above ships as a file you own — and a free 10-profile sample matched to your practice comes first.

A full delivery contains:

  • leads.xlsx — every field for all rows, ranked; segment sheets per role family and metro
  • leads.csv / leads.jsonl — the same data for CRM import and programmatic use
  • segments/ — ready cuts: evidence-heavy, recent-job-change, per-employer-class
  • README — field guide, scoring logic, the full funnel with named drop reasons

Each row is the complete data object — career, education, evidence, contacts, scoring. See what one contact looks like, field by field →

AI outreach agents can draft the first touch straight from these fields — referenced to each person's actual career, so the mail lands with a real person, not a template.

Order this cohort — the next batch is already computable.

Re-runs are deduplicated against everything previously delivered to you and can be narrowed by seniority band, degree geography or employer profile. Datasets $0.19–0.42 per contact; managed outreach and API delivery available.

Or just write — a consultation is free: we will map this cohort to your practice.

Outricher · Visa Talent Engine · recurring series · latest batch 2026-07 · MEASURED · Outricher data