USCIS requires you to satisfy at least 3 of 10 criteria, then survive a second test where the officer decides whether your full evidence package actually reflects extraordinary ability. Most petitions fail the second test, not the first.
The EB-1A is a two-step test established by the Ninth Circuit's 2010 decision in Kazarian v. USCIS. Step one: satisfy at least three of ten listed criteria under 8 C.F.R. § 204.5(h)(3). Step two: survive the final merits determination, where the adjudicator evaluates whether your evidence, considered as a whole, demonstrates that you have risen to the very top of your field. Most petitions are denied at step two, not step one. Understanding this distinction shapes every strategic decision.
For software engineers, ML researchers, and systems architects, criteria 3 (press), 4 (judging), 5 (contributions), 6 (publications), 8 (critical role), and 9 (high salary) are the most realistic targets. Criteria 7 and 10 are designed for artists and performers. Criteria 1 and 2 are possible but slow: most major tech awards run on annual nomination cycles, and relevant associations are few and selective.
If you work at a major technology company and earn above the 90th percentile for your occupation, you likely already satisfy criterion 9 without any new activity. Compare your compensation against Bureau of Labor Statistics Occupational Employment and Wage Statistics data for your SOC code. A staff engineer or senior researcher at a top-tier company earning above $250,000 in total compensation will clear this bar in most technical roles. Document it now and move on.
Salary is the only criterion you can satisfy retroactively with paperwork you already have. Every other criterion requires advance planning. Lock this one in first.
Peer review for academic conferences and journals satisfies the judging criterion under 8 C.F.R. § 204.5(h)(3)(iv) and is the fastest to activate. Most major venues, including NeurIPS, ICML, ICLR, OSDI, and SOSP, recruit reviewers through public signup portals before each submission cycle. A reviewer with 2 to 4 substantive review cycles at recognized venues, documented with invitation emails and a letter from the program chair confirming the selection criteria, can satisfy this criterion in under six months.
What USCIS is evaluating is not the number of reviews but whether you were selected because of demonstrated expertise. The documentation package must show that reviewers are chosen based on domain knowledge, not just availability. A one-paragraph letter from the program chair stating the selection basis is sufficient and carries significant evidentiary weight.
Original contributions of major significance is the criterion that carries the most weight in the final merits determination, and the one most applicants underestimate. The key word is major. A well-used open-source library with measurable adoption is a contribution. A widely cited technical report that influenced industry practice is a contribution. A feature shipped on a product used by millions is not, by itself, unless you can demonstrate that it changed how the field operates.
The strongest evidence for criterion 5 is citation data, adoption metrics, external commentary from recognized experts, and documented influence on subsequent work. If peers reference your methodology in their own papers or architectures, that is the evidence USCIS is looking for. Start collecting this documentation now, before you file.
Criterion 3 requires published material in major trade or professional publications. A blog post on your company website does not qualify. Coverage in The New York Times, Wired, MIT Technology Review, IEEE Spectrum, or a recognized trade journal in your domain does. The outlet must be recognized independently of you. A press placement in a publication that would cover anyone's announcement carries less weight than an in-depth profile or technical write-up that specifically discusses your work and its significance.
Three strong criteria with thorough documentation is far more persuasive than six weak criteria that barely cross the threshold. USCIS officers read hundreds of petitions. Depth signals that the evidence is real. Breadth without depth signals that someone assembled it to check boxes.
After clearing the numeric threshold, the officer steps back and asks whether your overall record reflects someone who belongs in the top of their field nationally or internationally. This is the Kazarian final merits determination. A petition filled with mid-tier conference reviews, blog coverage, and a high salary often clears three criteria but fails here because none of the evidence individually or collectively suggests top-tier standing.
The strongest petitions have a coherent narrative: a specific technical domain, a track record that shows growing recognition, and external validators who are themselves recognized in the field. If your petition reads like a curated story about exceptional work, it will pass. If it reads like a checklist, it may not.
Want us to map your profile against these criteria?