Artificial intelligence (AI) now runs in the daily operations of all kinds of companies, not only research labs. According to U.S. Census Bureau data collected through May 2026, 39.7% of businesses in the information sector and 33.9% in finance and insurance reported using AI.  

Data scientists and machine learning (ML) engineers are the people behind that work, and many of them trained or began their careers outside the United States. If you are one of them and you want to work or live in the U.S., more than one immigration pathway may apply, depending on your credentials, your employer situation, and your long-term plans. 

Which pathway fits you is less about your job title than about the kind of record you can document. An academic researcher, an industry engineer, a senior technical leader, and a founder can all work in data science, yet each typically carries very different evidence, and that difference is what points toward different immigration categories. This article walks through four common career profiles, the immigration pathways they point toward, what each pathway examines, and what we’ve seen in our own recent data science approvals. 

How Your Career Shapes Your Options 

A job title does not decide an immigration category. The better starting point is your record: what you do, how others recognize or rely on that work, how well you can document it, and what you plan to do in the United States. The same achievement can count differently depending on the standard it’s measured against. 

Industry Data Scientist or ML Engineer 

Industry work is judged mostly on what got built and what happened when other people used it, so the strongest documentation tends to be objective proof of that impact rather than a publication list. When the work you plan to continue in the U.S. has effects beyond a single employer, an EB-2 National Interest Waiver (NIW) is often worth reviewing. When your record also shows a high level of recognition or achievement, O-1A or EB-1A may fit. If your U.S. plans center on a specific job at a specific company, an employer-sponsored route such as H-1B may be the more natural fit. 

The record usually turns on a few questions: Can you document adoption, deployment, measurable results, patents, or technical ownership? Does the work affect only an internal employer function, or is there evidence of broader industry, research, commercial, or public relevance? And do independent sources, not just your own account, document that significance? 

Academic ML Researcher 

Researchers usually document their careers through published work and the way the field responds to it. That kind of record maps naturally onto the extraordinary-ability standards behind O-1A and EB-1A, once it reaches the level of recognition those categories require. A researcher can also pursue an EB-2 NIW, where the focus shifts from past acclaim to the prospective importance of the work planned in the United States. 

What matters most is what the publication record shows: Do citations, peer review, conference activity, and collaborations demonstrate real influence in the field, and has original research been used or recognized by others outside your own institution? That evidence can support an endeavor-based analysis, an extraordinary-ability analysis, or both. 

Senior Technical Leader 

Principals, staff-level engineers, directors, and technical leads may have few publications but a strong record of scope: responsibility for critical systems, decisions that affect infrastructure used at scale, and compensation that reflects that seniority. Those facts can support an O-1A or EB-1A analysis, and they can also help an EB-2 NIW petitioner show that the professional is well positioned to carry out a proposed endeavor. If the plan is tied to one U.S. employer, employer-sponsored categories can fit. 

The key question is whether the record can document a leading or critical role and explain why the organization or project carrying that role is itself distinguished or significant, backed by original contributions, high compensation, or external recognition. That, in turn, shapes whether the professional is better positioned for employer-tied status, an independent petition, or both. 

Startup or Entrepreneurial Data Professional 

Founders combine technical and business evidence, which is a strength if the record keeps the company’s achievements separate from the individual’s own contributions. From there, the analysis connects the founder’s technical role, leadership, and results to whatever the immigration category under review requires. 

Two questions usually decide how far that record goes: What technical contribution belongs to you as opposed to the company? And can product adoption, customers, patents, or market results be independently documented rather than asserted? The stronger the answer, the more your proposed U.S. work can be evaluated on its own implications rather than the commercial success of one company. 

What Each Pathway Examines 

Different categories look at different parts of a record, and each connects to data science work in its own way. Here is what each one is testing, and why it tends to be relevant for this profession. 

EB-2 NIW: A Self-Petition Green Card 

The EB-2 National Interest Waiver (NIW) is a green card that does not require an employer to sponsor you. First, you qualify for the underlying EB-2 category through an advanced degree or exceptional ability. The waiver then removes the job-offer and Program Electronic Review Management (PERM) labor-certification requirements. Because no job offer is needed, you can self-petition, though an employer may also file.  

U.S. Citizenship and Immigration Services (USCIS) reviews the waiver under the three-part test from Matter of Dhanasar: whether the proposed work has national importance, whether the professional is well positioned to advance it, and whether, on balance, waiving those requirements benefits the United States. USCIS decides this case by case. A profession, an economic benefit, or a single piece of evidence does not establish eligibility on its own. 

For data scientists and ML engineers, the relevance usually sits in the first prong. Many plan work whose value is not confined to one employer: a method, model, or application that others can adopt, or work in an area the government has identified as important. USCIS guidance recognizes the importance of progress in science and technology and lists certain areas of artificial intelligence among critical and emerging technologies. Whether that helps a specific petition depends on how the proposed endeavor advances the identified problem or technology, not on the AI label by itself. 

What a Recent Data Science EB-2 NIW Approval Reveals About the Proposed Endeavor 

Data science and technology EB-2 NIW petitions tend to be strongest when they tie the applicant’s work to a specific U.S. need. USCIS focuses on what the applicant plans to do, the direct impact of that work, and how it could affect a broader field or part of the economy. 

One of our recent approvals shows how this works in practice. Our client, a computer science professional, proposed to help U.S. small and medium-sized businesses and charities make better use of cloud technology, data engineering, and business intelligence. Many of these organizations struggle with secure data management, cloud migration, governance, and compliance, and they often lack the tools to turn their data into useful insights. The petition showed how his expertise would help solve these problems. This framing showed that his work would address common technology challenges faced by a whole category of U.S. organizations. 

That is exactly what the EB-2 NIW framework calls for. An endeavor can have national importance when its implications reach across a field. Adjudicators are also directed to evaluate the potential impact of the specific endeavor, not the importance of the applicant’s occupation in general. 

A strong petition for data scientists and ML engineers should clearly explain: 

1. The U.S. problem the work will address 

2. The technical method or application the applicant will use 

3. The broader outcomes the work is likely to produce 

Presented in this way, data science, machine learning, cloud computing, or AI become tools for advancing a clear national-interest goal. The supporting evidence should then show two things: the endeavor’s potential impact and the applicant’s ability to carry it out. 

See If You Qualify

Get Your Free EB-2 NIW Visa Profile Evaluation

Evaluate My Profile

EB-1A: A Self-Petition Green Card for Extraordinary Ability

EB-1A is a green card for individuals with extraordinary ability, shown either by a one-time major internationally recognized award or by meeting at least three of ten regulatory criteria, followed by a final review of the whole record. It allows self-petition, with no employer, job offer, or labor certification. If you are weighing this against the waiver, our EB-2 NIW vs. EB-1A comparison breaks down the difference. 

For data scientists and ML engineers, the criteria most often in play are original contributions of major significance, authorship of scholarly work, judging the work of others through peer review or program committees, a leading or critical role, and high compensation.  

What Approved Data Science EB-1A Cases Have in Common 

A strong EB-1A record shows how an applicant’s technical work has led to recognition and influence in the field over time. This matters because USCIS reviews EB-1A cases in two steps. First, it checks whether the evidence meets the regulatory criteria. Then it looks at the record to decide whether the applicant has sustained national or international acclaim and is among the small percentage at the very top of the field. 

In practice, the best evidence links technical work to objective signs of significance. Examples include: 

• A model or algorithm adopted by people beyond the team that built it 

• A data architecture deployed at meaningful scale 

• A method that other organizations have built into their own systems or practices 

• Research that has shaped later work in the field 

• Innovations with measurable results that led others to recognize or rely on the applicant’s expertise 

These accomplishments carry more weight when the record documents their scale, reach, results, and influence, backed by independent evidence. This is especially true in data science, where impact often shows up through commercial deployment, product integration, industry adoption, or operational use rather than through publications alone. Publications, judging, critical roles, original contributions, high compensation, and other evidence should work to show sustained distinction and influence in the field. 

See If You Qualify

Get Your Free EB-1A Visa Profile Evaluation

Evaluate My Profile

O-1A : A Temporary Visa for Extraordinary Ability 

The O-1A is a temporary visa for individuals with extraordinary ability in the sciences, business, education, or athletics. It draws on the same kind of evidence as EB-1A but is evaluated under eight criteria rather than ten, and a U.S. employer or agent must file it, so the individual cannot self-petition. It is not subject to an annual cap or lottery. 

For data scientists and ML engineers, the O-1A is often relevant as work authorization for someone with a recognized record who wants to avoid the H-1B cap, or as a step while building toward an EB-1A or another green card. The supporting evidence can come from academic, commercial, or industry achievements: influential research, significant technical contributions, peer review, critical roles for well-regarded organizations, recognized open-source projects, or patents. As with EB-1A, eligibility depends on how the evidence satisfies the regulatory criteria. 

What Successful Data Science O-1A Petitions Show About Top-of-Field Expertise 

O-1A petitions for data scientists and machine learning engineers can draw on a wide range of achievements, including technical, commercial, academic, and industry work. The key is showing sustained recognition in the field. Under USCIS regulations, the applicant must show a level of expertise that places them among the small percentage at the very top of their field.  

For data science professionals, technical work can be strong evidence of this standing when its significance is clear. That significance can come from adoption by others, deployment at scale, measurable results, industry use, or independent recognition. For example, a model, algorithm, platform, or data system that has been used across major operations, or that others rely on, helps show the reach and importance of the applicant’s work. 

Other evidence including critical roles, judging, publications, awards, and high compensation can all show how the field recognizes the applicant. For each accomplishment, the record should explain why it matters, what the applicant personally contributed, and what recognition followed. 

In the end, USCIS weighs all the evidence together to decide whether the applicant has the sustained acclaim and expertise the O-1A requires. This approach suits data science well, because distinction in the field is often best shown through the real-world adoption, influence, and measurable impact of technical work. 

See If You Qualify

Get your free O-1 visa evaluation today.

Evaluate My Profile

Employer-Sponsored Categories 

EB-2 PERM and the Employment-Based Third Preference (EB-3) are employer-sponsored green cards built around a permanent position and, in most cases, PERM labor certification; they tend to be relevant when a data scientist’s strongest basis is a long-term U.S. position rather than a portable endeavor. 

The L-1 Intracompany Transfer visa moves an employee within a qualifying multinational company and depends on the corporate relationship and prior employment abroad, not the job title. An ML engineer with specialized knowledge of a multinational’s models, platforms, architecture, or datasets may fit the L-1B category, while a data-science leader transferring into a managerial or executive role may fit L-1A.  

When your strongest basis is a specific U.S. job, the employer files. The H-1B covers specialty-occupation roles that require a bachelor’s degree in a related field. Data scientist and ML engineer positions commonly fit when the job duties and degree requirements meet that standard, though the analysis stays specific to the sponsored position. Many H-1B cases are subject to the annual cap, while universities and certain research organizations are cap-exempt, which is often relevant for academic ML researchers. 

Comparing the Options 

Each pathway has its own legal requirements, and a professional may have more than one option to weigh at the same time. 

Pathway	Status	Who files	What the record is built around
EB-2 NIW	Green card	You may self-petition; an employer may also file	Underlying EB-2 eligibility plus all three parts of the Dhanasar test
EB-1A	Green card	You may self-petition	Extraordinary ability, sustained recognition, and a final review of the whole record
O-1A	Temporary work	A U.S. employer or agent files	Extraordinary-ability evidence under the O-1A criteria; no annual cap or lottery
H-1B	Temporary work	A U.S. employer files	A specialty-occupation position and the worker’s qualifications
L-1A / L-1B	Temporary work	A qualifying U.S. organization files	A qualifying corporate relationship, prior work abroad, and managerial, executive, or specialized-knowledge role
EB-2 PERM / EB-3	Green card	A U.S. employer files	A permanent position, worker qualifications, and, in most cases, PERM labor certification

What We’re Seeing in Our Own Data Science Cases 

Across our cases we have seen industry records built on applied, cross-sector work approved under the EB-2 NIW, entrepreneurial records built around a product and a business approved under the same category, and research-heavy records approved under EB-1A. What the files share is not a job title or a degree, but a record documented specifically enough for the standard it was measured against.  

EB-2 NIW: Industry Data Scientist Working Across Sectors 

Colombo & Hurd secured an EB-2 NIW approval for a data scientist with a master’s degree in marketing analytics and a bachelor’s degree in computer engineering. His work spanned data science, AI, and ML applications in finance, healthcare, and education, and his proposed U.S. work included helping small and medium-sized businesses use data-driven tools. USCIS approved the EB-2 NIW petition. Read the full data scientist EB-2 NIW case study here.  

Why it is useful to you: This is an applied, industry record, not an academic one. If your value shows up in deployed work across sectors rather than in a publication list, this is the shape of a record that has met the EB-2 NIW standard. 

EB-2 NIW: Data Professional Building a Business Platform 

We also secured approval for a data analyst with a bachelor’s degree in computer information technology and a master’s degree in data science. His proposed work centered on an online platform that combines e-commerce, digital marketing, and data-analytics tools for small businesses. USCIS approved the EB-2 NIW petition, and he later became a permanent resident. Read the full data analyst EB-2 NIW case study here.  

Why it is useful to you: The record was built around a product and a business rather than a single employer or a research program. That is a common situation for entrepreneurial data professionals weighing the EB-2 NIW, and it shows the category can accommodate it. 

EB-1A: AI Researcher with Sustained Recognition 

We helped an AI and computer-science researcher from Mexico who had more than 23 years of experience, a Ph.D., two master’s degrees, more than 200 citations, scholarly publications, and judging activity. USCIS approved the EB-1A petition after a Request for Evidence (RFE) response. Read the full AI researcher EB-1A case study here.  

Why it is useful to you: This record rested on sustained recognition, not on a proposed endeavor. For an established researcher, it shows why it is worth comparing the EB-1A and EB-2 NIW standards rather than assuming that a strong technical record automatically fits one of them. 

Frequently Asked Questions 

Can open-source ML work support an immigration petition? 

Yes, when the evidence connects the work to the specific legal requirement under review. Independent adoption, use by others, technical influence, or recognition can all matter, and the category you are pursuing decides how that evidence is weighed. 

Do data scientists need academic publications for O-1A or EB-1A? 

No single type of evidence is required. Scholarly authorship can help, but the regulations also recognize original contributions, judging, leading or critical roles, awards, published material about the individual, and high compensation. The complete record has to meet the category’s standard. 

Can a data scientist self-petition for permanent residence? 

Yes, through the EB-2 NIW or EB-1A, if the requirements of that category are met. The two use different legal standards, so the right choice depends on which one your record fits. 

Can a data scientist pursue a green card while working in H-1B status? 

Yes. H-1B permits dual intent, so a worker may pursue permanent residence while keeping H-1B status. The immigrant petition still must satisfy the requirements of the green-card category chosen. 

Do data scientists who change employers often hurt their chances at a green card? 

Not for the self-petition categories. The EB-2 NIW and EB-1A are built around you and your work, not a specific employer, so moving between companies does not by itself weaken a case, and neither category ties you to one sponsor. Employer-sponsored routes such as H-1B, L-1, and EB-2 PERM are tied to a specific position, so job changes matter more there and can affect timing. Which pattern applies depends on the category and where you are in the process. 

Can proprietary work that cannot be published still count as evidence? 

Yes. Published work is only one kind of evidence. Confidential or proprietary work can be documented in other ways, such as adoption or reliance by others, measurable results, the scope of your responsibility, patents, or letters from people familiar with the work. What matters is whether the record shows the significance of the work under the standard being applied, not whether it appeared in a journal. How much any single item helps depends on the individual case. 

Do you need to work in a high-profile area like healthcare AI or national security to have a strong case? 

No subfield is required. The question is how your specific work connects to something of national importance, not whether it falls under a popular label. Work in a widely discussed area is not automatically stronger, and work outside one is not automatically weaker. That connection is assessed case by case, on the evidence in the individual petition. 

Next Steps 

Two data scientists with the same title can reach the same goal through completely different records. A useful review starts with your education, work history, evidence of impact or recognition, your proposed U.S. work, your employer situation, and your long-term plans, and from there identifies which categories deserve a closer look. 

Colombo & Hurd can review your background and explain which pathways fit your circumstances.