Adjudication-grade evidence mapping

EB-1A for Machine Learning Engineers

ML engineering evidence lives in latency numbers, cost reductions, and the open-source tools other teams adopt. Your record should show that you did not just deploy a model — you changed how models are shipped.

Built for: ML engineers, MLOps leads, and applied ML specialists who productionize models and build inference infrastructure.

Evidence matrix: your work mapped to USCIS criteria

Each row converts routine professional output into a packaged exhibit an adjudicator can score.

  • Daily work

    Designing model-serving or training infrastructure used across the organization

    USCIS criterion

    Original contributions of major significance

    Evidence packaging

    Architecture docs, adoption metrics, cost/latency improvements, and independent letters on how the system influenced practice.

  • Daily work

    Maintaining an open-source ML library, compiler, or serving framework

    USCIS criterion

    Original contributions of major significance

    Evidence packaging

    Download and dependency data, named production users, and expert letters explaining the technical novelty.

  • Daily work

    Publishing papers or technical blog posts on ML systems

    USCIS criterion

    Authorship of scholarly articles / published material

    Evidence packaging

    Publication record, venue circulation, and citation or reference trails from other practitioners.

  • Daily work

    Reviewing ML systems papers or workshop submissions

    USCIS criterion

    Judging the work of others

    Evidence packaging

    Reviewer invitations, completed reviews, and program-chair confirmation of selection standards.

  • Daily work

    Leading the applied ML team or platform function

    USCIS criterion

    Critical role for distinguished organizations

    Evidence packaging

    Org placement, team size, budget, and outcomes owned such as model count, serving reliability, or inference cost.

  • Daily work

    Top-decile compensation for ML engineers in your metro

    USCIS criterion

    High remuneration

    Evidence packaging

    Total-comp statements benchmarked against BLS/OES and industry survey data for ML engineers in the same location.

High remuneration benchmarks

Indicative BLS/OES-style 90th percentile total compensation for this occupation by metro. Your petition must compare your compensation against the correct occupation code and location.

$287,000

Bay Area — 90th percentile

$258,000

New York City — 90th percentile

$271,000

Seattle — 90th percentile

Calculate Your Salary vs 90th Percentile OES

RFE mitigation for this profile

Where USCIS pushes back

ML engineers often face the 'implementer, not inventor' objection. USCIS concedes the system works at scale but questions whether the design was original to you. Mitigate by isolating the components you architected, documenting external adoption, and securing letters from users outside your company who built on your tooling or patterns.

Frequently asked questions

Related visa guides

See where your profile stands today

Run the free profile audit and get a criterion-by-criterion read on your EB-1A eligibility, tuned to your role.