NIW Bearings
remandedNOV062025_01B52032025-11-06 · appeal

This is an AAO appeal decision — a case denied once and appealed. It shows how the framework is applied; it is not the filing population.

The petitioner, a PhD student in computer science, proposed to enhance AI model training and inference efficiency, in-context learning, and collaborative intelligence frameworks, with an eye toward eventually pursuing a professorship.

Field: artificial intelligence / computer science (LLM training and inference research) · Read the decision (uscis.gov)

EB-2 threshold — not met

SCOPS never adjudicated exceptional-ability eligibility; AAO found the record as constituted did not sufficiently establish it and remanded for SCOPS to make findings on the claimed three criteria and conduct a final merits determination.

Prong 1 — national importanceestablished

We agree with the Petitioner that SCOPS erred in identifying the proposed endeavor as a tenure track assistant professor.
the importance of an industry or field is not sufficient to meet the national importance requirement under the Dhanasar framework if the evidence submitted does not specifically discuss the Petitioner's research and its impact.

AAO decision text

How the evidence was treated

  • recommendation letter · ignored
  • certification membership · ignored
  • citations publications · ignored
  • degree · ignored
  • resume experience · discounted
    a graduate student with relatively brief periods of research experience prior to filing the petition
    AAO decision text

This record is one of thousands, each coded for the reasons it turned. A placement shows where your profile sits in that record, not what it predicts. See where your profile sits — 90 seconds

Notable

Rather than affirming or reversing, the AAO remanded because SCOPS never made findings on underlying EB-2 exceptional-ability eligibility or on Dhanasar prongs 2 and 3, and mischaracterized the endeavor (labeling it 'business industry' and reducing it to a 'Tenure-Track Assistant Professor' job title) without engaging the petitioner's actual AI research claims. The AAO also flagged an apparent signature discrepancy between the appeal filings and the initial petition (footnote 1) without further comment.

Authorities this decision leans on

From the doctrinal survivor set — 187 authority tests across 47 distinct authorities cleared the differential-lift gates. Only those appear here.

  • Flores v. Garland
  • USCIS Policy Manual F.5(B)(2)
  • USCIS Policy Manual F.5(D)(2)

What this case teaches

Analyst reading of the decision text.

Decisive factor

SCOPS mischaracterized the endeavor as a job title ('Tenure-Track Assistant Professor'/'business industry') instead of the AI research itself, and failed to analyze specific record evidence, precluding meaningful review; case remanded for proper Dhanasar and exceptional-ability analysis.

Transferable lesson

Frame the endeavor as the specific research/technical undertaking (not the job title) and require adjudicators—and petitioners—to tie concrete evidence to that endeavor's scaled national impact, not just field importance.

Endeavor framing

moderate

National-importance theory

critical_emerging_tech · field_advancement · stem_positive_factor

Evidence targeting

mixed

Cases in adjacent profiles