You Don't Need a Nobel Prize to Get the O-1 Visa. Here's How.

A senior engineer came to me last year with seven years of experience at two FAANG companies, three open source projects on GitHub, and a total compensation package in the top 5% of his field. He had lost the H-1B lottery twice. His first question was: "Do I even qualify for the O-1?"

He qualified for five of the eight criteria. We filed. He was approved without an RFE.

I tell this story because it reflects something I see constantly in my practice. Talented engineers underestimate their own profiles. They look at the phrase "extraordinary ability" and assume it means Nobel Prize or Olympic medal. It does not. What it means, in practice, is a career record that is documentable, quantifiable, and meaningfully above the norm in your field. For software engineers and AI researchers, that record often lives on GitHub, LinkedIn, Google Scholar, and your pay stub.

Let me show you exactly how I read it.

What "Extraordinary Ability" Actually Means for a Software Engineer

The O-1A visa requires you to demonstrate extraordinary ability in science, education, business, or athletics. Software engineering qualifies. AI research qualifies.

You need to meet at least three of eight criteria defined by USCIS, or provide comparable evidence where standard criteria do not fit your discipline. USCIS adjudicated 31,681 O-1 petitions in fiscal year 2025, approving 29,733 of them. That is a 93.9% approval rate. The category is not as hard to crack as people think.

What separates approved petitions from denied ones is almost always the same thing: whether the evidence file was built carefully and with specificity, or assembled in a hurry with vague language and generic letters.

The eight criteria are:

Major awards
Selective memberships
Published material about you
Judging the work of others
Original contributions of major significance
Authorship of scholarly articles
Critical role at a distinguished organization
High remuneration

You need three. Most strong senior engineers qualify for four or five without realizing it.

Pull up your LinkedIn, GitHub, and most recent offer letter. Keep them open. I will walk you through each criterion that matters for tech professionals.

Criterion 1: Original Contributions — This Is Where GitHub Lives

This is the most powerful criterion for software engineers and the most misunderstood.

USCIS wants evidence of "original contributions of major significance" in your field. For an engineer, that translates directly to your open source work, your patents, your internal systems used at scale, and your documented product impact.

Here is what I look for when a client sends me their GitHub profile:

Stars and forks. A project with 10,000 or more stars, particularly with documented adoption by named major companies, satisfies the original contributions criterion. Stars alone are not enough. What matters is the combination: thousands of peer developers finding your work valuable, plus named organizations using it in production.
Corporate adoption. If Google, Microsoft, Stripe, or any recognizable company has publicly acknowledged using your code, that is evidence. Blog posts, conference talks, dependency graphs, and engineering letters from teams that deployed your work all count.
Maintainer status. Contributing to highly selective projects like the Linux kernel, Chromium, or TensorFlow, where only a small percentage of contributions are accepted, proves extraordinary ability through the selection process itself. If you are a core contributor or maintainer, document it.

What does not work: a GitHub with dozens of small personal projects and no adoption evidence. Volume is not impacted. USCIS is looking for significance, not activity.

Identify your top one to three projects by stars, forks, and named corporate adoption. That is the core of your original contributions.

Criterion 2: Judging — Your Code Reviews Count

Most engineers skip this criterion because they assume judging means sitting on an academic panel. It does not.

USCIS defines judging as evaluating the work of others in your field, either individually or as part of a panel. For a software engineer, this includes:

Peer code reviews with documented history on GitHub or internal systems
Technical hiring panels and final-round interview participation
RFC reviews and standards committee participation
Editorial roles in technical publications

Here is how I document it. I ask clients to pull their code review history. I want to see the volume, the seniority of the engineers whose work they reviewed, and the complexity of what they evaluated. A staff engineer at Meta who has reviewed 500 pull requests from senior engineers has strong evidence for this criterion.

For AI researchers, peer review of conference submissions at NeurIPS, ICML, CVPR, or ICLR is clean, direct evidence. If you have reviewed papers, request confirmation letters from the program committees.

Email your engineering manager this week and ask for a letter documenting your code review history and any technical hiring panel participation. That letter may cover this criterion entirely.

Criterion 3: High Remuneration — Your Offer Letter Is Evidence

This criterion is more straightforward than the others, and one of the easiest for senior tech professionals to satisfy. You must show that you command a high salary compared to others in your field.

In practice, I look for total compensation in the top 10 to 25% of your role, level, and geography. Here is how I benchmark it:

Levels.fyi for total compensation by role, level, and company
Bureau of Labor Statistics for occupational wage data by geography
DOL H-1B salary data which is publicly searchable and shows prevailing wages by role

If you are a senior or staff engineer at a major tech company with total compensation including base, equity, and bonus, you almost certainly satisfy this criterion. Document it with your offer letter, most recent pay stub, and a comparator showing where your compensation sits relative to peers.

For engineers at earlier-stage startups, this criterion may be harder to satisfy. In those cases, I look at other criteria first and use remuneration as a supporting element.

Go to Levels.fyi and find the median total compensation for your exact role, level, and location. If your compensation is in the top 15 to 20%, document it.

Criterion 4: Critical Role — This Is What Your LinkedIn Actually Shows

Being employed in a critical or essential capacity at a distinguished organization is one of the most flexible criteria for tech professionals, and one of the most commonly mishandled.

Two things must be true here:

Distinguished organization. USCIS wants your employer to have a recognized reputation in the field. This is straightforward for FAANG, large fintech companies, or well-funded startups with press coverage. For lesser-known companies, we need to build the distinguished reputation evidence through press, funding announcements, and industry recognition.
Critical capacity. I need evidence that your role was essential to the organization's operation, not just that you were a good employee. Performance reviews, engineering letters that speak to your indispensability, and evidence of technical systems you owned that generated significant revenue or user growth all work here.

What does not work: a title of "Senior Software Engineer" with no supporting narrative. Titles alone tell USCIS nothing.

Ask your VP of Engineering for a letter describing what would have been different at the company if you had not been there. That specificity is what USCIS responds to.

What AI Researchers Need to Know Specifically

If you are an AI researcher, your evidence landscape looks different from a software engineer's, and in many ways stronger.

Here is how the criteria map naturally for a researcher:

Published papers at tier-one conferences (NeurIPS, ICML, CVPR, ICLR) satisfy the scholarly articles criterion
Citations satisfy original contributions
Peer review assignments satisfy judging
High post-PhD industry compensation satisfies remuneration
Named model or dataset adoption satisfies original contributions through comparable evidence

The one area where AI researchers often struggle is the critical role criterion. If you are at a university or a research lab without a clear commercial product, we need to establish the lab's reputation through funding sources (DARPA, NSF, NIH grants carry significant weight) and downstream industry impact of the research.

One important note: approximately 15 to 20% of O-1 petitions receive a Request for Evidence, and of those, around 60 to 70% are ultimately approved after the RFE response. An RFE is not a denial. But building the petition correctly the first time is always better than responding to an RFE under pressure.

Pull your Google Scholar profile and note your citation count and h-index. Check whether any conference papers have been cited by industry papers or used in commercial systems. That is the evidence I need.

How to Know If You Are Ready to File

After this walkthrough, here is a simple self-assessment. Be honest with yourself.

You are likely ready to file if you meet three or more of these:

A GitHub project with 5,000 or more stars and named corporate adoption
Substantial peer code review or technical hiring at a recognized organization
Total compensation in the top 15% of your role and geography
A critical role at a FAANG company or a well-funded startup with press coverage
Published work at a top-tier conference or in a recognized technical publication
Core contributor status on a major open source project

You are borderline if:

You meet two criteria confidently and one weakly
Your organization is not well known and we need to build the distinguished reputation evidence

This is fixable, but it takes planning.

You are not ready yet if:

You have fewer than five years of experience with limited external recognition
No open source footprint, no publications, no external press

The right move in that case is to start building the evidence file now, not wait until the H-1B lottery fails again.

Go to Visa Architect and book a profile evaluation. I will tell you honestly where you stand, which criteria you already satisfy, and what we need to build before filing.

You Have More Evidence Than You Think. Let's Find It.

Most engineers who come to me for a first consultation leave with a clearer picture than they expected. Not always a green light, but always a specific roadmap. Here is what I need. Here is what we are building. Here is how long it realistically takes.

If you are a software engineer or AI researcher who has lost the H-1B lottery, is running out of OPT time, or is simply tired of waiting for a system that was not built for people like you, the O-1 may already be within reach.

The first step is knowing where you actually stand, not where you assume you stand.

At Visa Architect, we review profiles personally and give you the honest answer. Not a sales pitch. Not a guarantee. A clear-eyed assessment of your case and what it would take to win.

Book your profile evaluation at visarchitect.com

FAQ

Do I qualify for the O-1 visa as a software engineer if I do not have a PhD?

Yes. The O-1A does not require any educational credential. The standard is extraordinary ability demonstrated through documented achievement. Software engineers qualify based on open source adoption, product impact, compensation, and technical leadership, none of which require a degree. Many strong O-1 engineers I have filed for are self-taught or hold bachelor's degrees only.

How many GitHub stars do I need for O-1 visa evidence?

There is no official USCIS threshold. In practice, projects with 5,000 or more stars combined with documented corporate adoption make strong evidence for the original contributions criterion. Stars alone are not enough. What matters is the combination of peer recognition and independent adoption by named organizations with documented production use.

Can I file an O-1 visa without a job offer?

No. The O-1 requires a U.S. sponsor, which can be an employer or a licensed agent. You cannot self-petition. However, the sponsor does not need to be a traditional employer. You can work through an agent who represents you across multiple clients, which is a structure I use for founders and freelance engineers with multiple U.S. work relationships.

How long does O-1 visa processing take in 2026?

As of October 2025, USCIS reports 80% of O-1 cases finish within 7 months, or 15 business days with premium processing. I recommend premium processing for most clients. The predictability is worth the fee, especially if you have an employment start date or a visa expiration to manage.

What is the difference between the O-1 visa and EB-1A green card for software engineers?

The O-1 is a temporary work visa. The EB-1A is a permanent green card. Both use the extraordinary ability standard, but the EB-1A bar is higher and USCIS applies a more rigorous final merits review. Many of my clients file for the O-1 first to get work authorization immediately, then build a stronger evidence file for the EB-1A while working in the U.S. This is one of the cleanest routes to a green card for high-skill tech professionals, especially for Indian nationals stuck in the EB-2 backlog.