LAW OFFICE OF HONG-MIN JUN

Law Office of Hong-min Jun · Chicago · Indianapolis

Menu
HomeBlog
Insightful ArticleApril 29, 2026

NIW for Software Engineers and AI Researchers: The Framework That Actually Works

Software engineers and AI researchers are structurally well-positioned for NIW — but only when the petition is built around the correct framework. Attorney Hong-min Jun explains the three structural errors that sink most technology NIW petitions, the six national interest categories most defensible in 2026, and the evidence architecture that works for both academic researchers and industry engineers.

J
Attorney Hong-min Jun
Law Office of Hong-min Jun P.C.
LinkedIn

Key Takeaway

Software engineers and AI researchers are structurally well-positioned for NIW — but only when the petition is built around the correct framework. Most technology professionals fail because they treat NIW as a credential review. This article covers 3 structural errors, 6 national interest categories, and 4 real approval case patterns.

Why Technology Professionals Are Structurally Advantaged

Software engineers and AI researchers occupy a structurally favorable position in the NIW framework — not because USCIS gives preferential treatment to the technology sector, but because the evidentiary requirements of the Dhanasar three-prong test align naturally with how technology work is documented, deployed, and measured.

Unlike performing artists or business consultants, technology professionals produce independently verifiable artifacts.

Code & Systems
Code repositories, deployed systems, published models
Intellectual Property
Patents, peer-reviewed papers, citation records
Adoption Metrics
Institutional adoption, open-source contribution data

The challenge is not a shortage of evidence — it is the failure to frame that evidence correctly. Most technology professionals who attempt NIW without experienced counsel make one of three structural errors.

Part I — The Three Structural Errors

01

The Credential Review Petition

The most common mistake: structuring the petition as a comprehensive listing of degrees, publications, GitHub repositories, citation counts, and employment history — as if the volume of credentials itself establishes NIW eligibility.

Core misunderstanding: The Dhanasar framework does not ask how impressive your credentials are. The adjudicator is not looking for a resume — they are looking for a structured legal argument.

02

The Generic National Importance Claim

Asserting national importance without connecting it to a specific, documented U.S. need.

✗ Fails

“AI will transform the U.S. economy”
“Software infrastructure is critical to national security”

✓ Works

“My endeavor addresses the documented shortage of AI safety researchers in U.S. federal agencies, as identified in [specific NIST report]”

03

The Vague Proposed Endeavor

A proposed endeavor statement too broad to be evaluated. A defensible proposed endeavor must specify all five of the following.

  • Technical domain (what field)
  • Specific problem being addressed (what problem)
  • Methodology or approach (how it is solved)
  • Target deployment context (where it is applied)
  • Measurable outcomes indicating success (how it is measured)

Part II — Six National Interest Categories for 2026

Each category is grounded in specific federal policy documentation. Identify which category your work falls under.

1

AI Safety & Alignment Research

Alignment techniques, interpretability research, robustness evaluation, LLM red-teaming methodologies

Policy basis: Executive Order 14110, NIST AI Risk Management Framework, OSTP National AI R&D Strategic Plan
2

Critical Infrastructure Cybersecurity

Power grids, water systems, financial systems, healthcare networks, transportation security

Policy basis: CISA annual vulnerability reports, National Cybersecurity Strategy (2023), 700,000+ unfilled positions
3

Healthcare AI & Medical Technology

Early disease detection, clinical NLP, hospital resource allocation, rural AI diagnostics

Policy basis: NIH NCATS, FDA Digital Health Center of Excellence, CMS Innovation Center, HRSA shortage designations
4

Clean Energy & Climate Technology

Grid optimization, energy storage management, renewable energy predictive maintenance, carbon capture

Policy basis: Inflation Reduction Act (2022), Infrastructure Investment and Jobs Act (2021), DOE OSTI reports
5

National Security & Defense Technology

Autonomous systems, signals intelligence, military logistics optimization, defense network security

Note: Classified work cannot be disclosed. Build argument around public policy documents (National Defense Strategy, DoD AI Strategy)
6

Semiconductor & Hardware-Software Co-Design

Compiler optimization, chip design automation, AI accelerator development

Policy basis: CHIPS and Science Act (2022), NIST CHIPS Program reports, Department of Commerce semiconductor capability reports

Part III — Evidence Architecture

① The Proposed Endeavor Statement — Most Critical Document

Must be technically specific without appearing to describe a single job position. The goal is to describe a research or development program broader than any single employer's needs, but specific enough to be evaluated against the national importance standard.

✗ Weak (Fails)

“I will develop AI systems that improve healthcare outcomes in the United States.”

✓ Strong (Approved)

“I will develop and deploy federated learning architectures for clinical decision support systems in rural critical access hospitals, enabling AI-assisted diagnostic capabilities without requiring patient data to leave local hospital networks. This approach addresses the documented shortage of diagnostic specialists in rural U.S. healthcare settings, as identified in the HRSA Health Professional Shortage Area designations, while complying with HIPAA data residency requirements.”

② Evidence Type Strategy

Publication & Citation

Present by relevance to proposed endeavor, not raw count. 50 citations on relevant papers > 500 on unrelated subfields

Deployment & Adoption

Primary Prong Two evidence for engineers without publications. Production metrics, institutional adoption, GitHub stats, org letters

Patent Evidence

Independently verifiable, proves novelty & non-obviousness. Must connect technical claims to national interest category

Key insight: A software engineer whose work has been deployed in 50 U.S. hospital systems is making a stronger national importance argument than a researcher whose papers have been cited 200 times. Deployment evidence demonstrates real-world impact directly.

③ Expert Letters — Who to Ask

✓ Persuasive Letters
  • Senior researchers at national labs (NIST, NIH, DOE)
  • Program managers at federal agencies (NSF, DARPA, ARPA-E)
  • CTO / Technical Director at orgs that deployed your work
  • Independent academic experts
✗ Weak Letters
  • Colleagues or supervisors at current employer
  • Co-authors (lack independence)
  • Letters addressing only technical quality
  • Letters that never mention national importance

Part IV — Four Approval Case Patterns

Case 01 AI Safety Researcher — Chinese National, CS Ph.D.
Profile

Ph.D. from top-10 U.S. program, 12 papers on LLM interpretability & alignment, 340 citations

Proposed Endeavor

Evaluation frameworks for detecting deceptive alignment in AI systems deployed in high-stakes contexts (criminal justice, healthcare triage, financial credit)

Key Evidence

Executive Order 14110, NIST AI Risk Management Framework, letters from NIST researcher + NSF program officer + AI safety org senior researcher

Why It Worked

Specific proposed endeavor + federal policy documents directly confirming national priority + publication record directly tied to the endeavor

Case 02 Healthcare AI Engineer — Korean National, No Publications
Profile

M.S. in CS, 6 years at major healthcare tech company, zero academic publications. Deployed systems used by 200+ U.S. hospitals, reducing clinical documentation by 40 min/physician/day

Proposed Endeavor

Federated learning infrastructure for rural critical access hospitals — enabling AI-assisted clinical decision support without patient data leaving local networks

Key Evidence

HRSA shortage area data, CMS rural health reports, letters of intent from 3 rural hospital systems committing to pilot the proposed system

Why It Worked

Approved with zero publications. Strong deployment evidence + specific endeavor + hospital letters of intent (implementation evidence) was decisive

Case 03 Semiconductor Software Engineer — Indian National, 4 Patents
Profile

Ph.D. in EE, 8 years developing compiler optimization tools for AI accelerator chips, 4 patents on novel compilation techniques, open-source compiler contributions used by major semiconductor companies

Proposed Endeavor

Open-source compiler toolchains for domestic AI accelerator chips under the CHIPS Act — reducing U.S. dependence on foreign compiler infrastructure

Key Evidence

CHIPS and Science Act, NIST CHIPS Program reports, letters from 2 U.S. semiconductor companies confirming importance to domestic manufacturing programs

Why It Worked

CHIPS Act provided exceptionally strong legislative support. Patent record + open-source contributions served as strong Prong Two evidence

Case 04 Cybersecurity Researcher — Ukrainian National, 23 CVEs
Profile

M.S. in CS, 7 years vulnerability research on U.S. power grid industrial control systems, 23 CVEs disclosed, DEF CON & Black Hat presenter, CISA coordinated disclosure collaborator

Proposed Endeavor

Automated vulnerability detection tools for legacy industrial control systems in U.S. power generation facilities — addressing national security risks in unpatchable systems

Key Evidence

CISA critical infrastructure reports, National Cybersecurity Strategy, direct CISA letter confirming national security priority

Why It Worked

The CISA letter was decisive. A federal agency directly confirming national security priority is among the strongest possible NIW evidence

Part V — EB-1A vs. NIW Decision Guide

EB-1A Extraordinary Ability
  • Sustained national or international acclaim required
  • Major awards, high citation counts, invited keynotes
  • No proposed endeavor needed — based on past achievement
  • Must demonstrate top tier of technical community
NIW National Interest Waiver
  • Compelling proposed endeavor required
  • Viable for engineers without academic publications
  • Suitable for researchers in emerging fields
  • Suitable for deployment-based professionals

Optimal strategy: Technology professionals with strong credentials and a compelling proposed endeavor should consider filing both simultaneously. The two petitions can be filed concurrently, and approval of either provides a path to permanent residency.

Part VI — Special Considerations for AI Researchers in 2026

The Rapidly Evolving Policy Landscape

The federal policy landscape for AI has evolved rapidly since 2023. Executive Order 14110, the NIST AI Risk Management Framework, and the AI Safety Institute's ongoing work have created a rich documentary foundation. A petition filed in 2026 should cite the most current versions of relevant policy documents — earlier versions may have been updated or superseded.

Open-Source Contributions

Many AI researchers and software engineers have significant open-source contributions not reflected in traditional academic metrics. The key is connecting them to the national importance argument: not just “my library has 50,000 GitHub stars” but “my library is used by [specific federal agencies, national laboratories, healthcare systems] for [specific applications] that address [specific national need].”

Industry vs. Academia

AI researchers at industry labs (Google DeepMind, OpenAI, Anthropic, Meta AI, Microsoft Research) often have publication records comparable to academic researchers, plus commercial deployment at scale. This dual nature is an advantage in NIW petitions — it allows both citation-based Prong Two arguments and deployment-based national importance arguments simultaneously. Do not downplay commercial work in NIW petitions.

Conclusion

Software engineers and AI researchers are structurally well-positioned for NIW — but only when the petition is built around the correct framework. The Dhanasar three-prong test rewards specificity, documentation, and forward-looking framing.

Technology professionals who treat NIW as a credential review will fail. Those who build their petition around a specific proposed endeavor connected to a documented national interest, supported by evidence that directly addresses each Dhanasar prong, have a strong pathway to approval.

The federal policy landscape in 2026 provides an exceptionally rich documentary foundation. AI safety, cybersecurity, healthcare AI, clean energy technology, and semiconductor software are all areas where federal policy documents explicitly identify national priorities that align with the work of software engineers and AI researchers. The framework is available. The documentation exists. The work is in building the argument correctly.

Attorney Note

This article reflects general legal analysis and does not constitute legal advice for any specific case. NIW eligibility depends on the specific facts of each applicant's situation. If you are a software engineer or AI researcher considering NIW, consult with an immigration attorney experienced in technology NIW petitions before filing.

Insights · More Articles

Related Posts

All Articles
Go Deeper

Related Resources on This Site

Explore our in-depth guides, analysis tools, and case studies related to this article.

Browse by Topic

Talk with Us