Success Stories
From Denial to Approval: robotics engineer's O-1 Journey — February 2025
Detailed analysis with practical recommendations for O-1 applicants at every stage.
The Initial Denial: Diagnosing What Went Wrong
This article traces the O-1A petition journey of a robotics engineer — referred to here as the beneficiary — whose initial petition was denied by USCIS in late 2024 and whose second petition was approved in February 2025 following a comprehensive record rebuild. The case illustrates the practical application of 8 CFR 214.2(o)(3)(iii)(B) criteria in the STEM context, the Kazarian step-two failure mode, and the specific evidentiary strategies that ultimately produced an approval. It is presented as a case study to inform the practice of other robotics and engineering practitioners navigating similar challenges.
The initial petition claimed four criteria: original contributions of major significance, judging the work of others, scholarly articles in major media, and high salary relative to others in the field. The denial notice acknowledged that the petitioner had submitted evidence for each criterion but found at step two that the totality of the evidence did not establish extraordinary ability. Specifically, the adjudicator found that: the scholarly article evidence consisted of papers in conference proceedings that were not sufficiently established as major media in the field; the judging criterion was supported by a single conference review assignment without documentation of the conference's significance; and the salary evidence compared the beneficiary's compensation to a broad engineering occupational category rather than to robotics specialists specifically.
Importantly, the denial did not challenge the quality of the beneficiary's actual work. The beneficiary had a legitimate and impressive research record — IEEE publications, a portfolio of granted and pending patents, and a GitHub presence with substantial open-source contributions that had been forked and cited by industry developers worldwide. The problem was not the credentials; it was the documentation and framing of those credentials in a petition that had been assembled quickly without the depth of evidentiary development the Kazarian standard demands.
IEEE Publications: Rebuilding the Scholarly Article Criterion
The scholarly article criterion under 8 CFR 214.2(o)(3)(iii)(B)(4) requires publications in professional journals or other major media. For the robotics engineer, the initial petition had relied primarily on publications in IEEE conference proceedings — technically peer-reviewed, but presented without context establishing those specific conferences as major venues in the robotics research community. The denial challenged whether conference proceedings constitute 'professional journals or other major media' without additional context, citing USCIS policy guidance on the distinction between journal publications and conference papers.
For the second filing, counsel undertook a systematic analysis of the beneficiary's publication record and identified publications in three IEEE journals: IEEE Transactions on Robotics, IEEE Robotics and Automation Letters, and the International Journal of Robotics Research. These journals publish peer-reviewed archival research (as distinct from conference proceedings) and are widely recognized as the leading publication venues in their field. The second petition documented each journal's impact factor, editorial board composition, acceptance rate, and standing in the robotics research community through declarations from independent expert witnesses and through published journal ranking data from Scimago and similar bibliometric sources.
For the conference proceedings that remained in the record, counsel added context establishing that the specific conferences — ICRA (International Conference on Robotics and Automation) and IROS (International Conference on Intelligent Robots and Systems) — are among the highest-impact publication venues in robotics research, with acceptance rates below 30% and citations per paper that rival or exceed those of many archival journals. Under 8 CFR 214.2(o)(3)(iv)(B), the adjudicator considering the totality of the evidence must evaluate proceedings publications in context, and that context was now fully developed in the record rather than left implicit.
Patent Portfolio Rebuild: From Listing to Demonstrating Impact
The beneficiary held seven granted U.S. patents and fourteen pending patent applications in robotic perception, manipulation, and human-robot interaction. The initial petition listed these patents in a brief exhibit with no analysis of their technological significance, commercial deployment, or citation by other inventors. The denial did not specifically reject the patent evidence, but it contributed nothing meaningful to the step-two analysis because it was presented as a bare inventory rather than as evidence of original contributions of major significance under 8 CFR 214.2(o)(3)(iii)(B)(5).
For the second filing, counsel developed the patent portfolio into a comprehensive original contributions exhibit. Each granted patent was analyzed against three dimensions: the technical problem it addressed, the novelty of the approach compared to prior art, and the evidence of commercial deployment or industry adoption. For the most significant patents — two in robotic grasping systems that had been licensed to three major robotics manufacturers — counsel obtained licensing agreements (redacted for commercial sensitivity) and letters from the licensees describing how the patented technology had been integrated into production systems. These materials transformed the patent listing from a credential inventory into evidence of real-world impact that directly addressed the major significance component of the criterion.
Patent citation analysis was added to the exhibit package: a professional patent analytics report documenting forward citations to the beneficiary's granted patents by subsequent inventors. The beneficiary's two most-cited patents had each accumulated over 50 forward citations within five years of grant, a metric that places them in the top tier of patent impact in their technology classification. This citation data provided the kind of objective, third-party validation of significance that the Kazarian step-two analysis values — not the beneficiary's self-assessment of their contributions, but measurable evidence of how the field had engaged with and built upon that work.
GitHub and Open-Source Contributions as Comparable Evidence
One of the most innovative elements of the second petition was the inclusion of open-source software contributions as comparable evidence under the O-1A framework. The beneficiary had developed and maintained several open-source robotics software libraries on GitHub, with aggregate repositories garnering over 8,000 stars, 1,200 forks, and contributions from developers at major robotics companies, research universities, and technology firms worldwide. This open-source activity did not fit neatly into any of the enumerated O-1A criteria but was relevant to the original contributions analysis and to the step-two totality-of-evidence determination.
Under USCIS policy, O-1A petitioners may submit comparable evidence where the standard criteria do not readily apply to the beneficiary's occupation. 8 CFR 214.2(o)(3)(iii)(B) acknowledges that not all fields map perfectly onto the enumerated criteria, and USCIS's 2022 policy guidance on O-1A adjudications specifically recognizes that STEM professionals in newer specialties may need to use comparable evidence to capture the full scope of their achievements. For the beneficiary's counsel, the GitHub and open-source contributions fit this framework — they demonstrated original technical contributions, peer adoption and engagement, and field-wide recognition that corroborated the criterion-based evidence in the record.
The comparable evidence exhibit for open-source contributions included: repository star and fork counts with screenshots and third-party analytics; a declaration from the beneficiary explaining the technical significance of each library; and letters from developers at industry-recognized companies describing how they had integrated the libraries into production robotics systems. One letter came from a senior engineer at a major autonomous vehicle company describing how the beneficiary's depth-estimation library had been adapted for use in the company's perception stack. This level of specificity — technical content, real-world deployment, recognized industry context — met the evidentiary standard for comparable evidence supporting major significance under 8 CFR 214.2(o)(3)(iii)(B)(5).
Salary Survey Strategy: Precision Comparisons for Robotics Specialists
The initial petition's salary evidence had compared the beneficiary's compensation to the broad BLS occupational category for computer and information research scientists, which includes a wide range of specialties at varying compensation levels. The adjudicator did not deny on the salary criterion specifically, but the broad comparison weakened the step-two analysis by failing to demonstrate that the beneficiary's salary placed them at the top of their specific peer group.
For the second filing, counsel developed a targeted salary comparison that benchmarked the beneficiary's total compensation (base salary, bonuses, and equity) against robotics-specific salary data from three independent sources: Levels.fyi compensation data for robotics software and hardware roles at technology companies, a survey of robotics engineer salaries at autonomous vehicle and robotics startup companies from a specialized technology compensation firm, and a declaration from an independent human resources consultant with experience in STEM compensation analysis who certified that the beneficiary's compensation placed them in the top decile of robotics engineers with comparable experience in comparable roles in comparable markets.
The multi-source approach addressed a known RFE pattern at the Vermont Service Center, where adjudicators have questioned salary comparisons based on a single survey or a broad occupational category without field-specific granularity. By providing three independent data points that consistently positioned the beneficiary's compensation above the 90th percentile for their specific discipline, the second petition gave the adjudicator the precision comparison needed to credit the salary criterion at step one and to weigh it appropriately in the step-two totality analysis under 8 CFR 214.2(o)(3)(iv)(B).
The Second-Filing Approach: Structural Differences That Mattered
The structural differences between the first and second petitions were as important as the evidentiary additions. The first petition's support letter ran 18 pages and addressed each criterion in a single paragraph with minimal citation to exhibits. The second petition's support letter ran 42 pages and was organized in two parts: a section-by-section criterion analysis with detailed exhibit references, and a narrative section presenting the beneficiary's contributions in the context of the field's development. The narrative section explained how the beneficiary's work in robotic perception had contributed to a transition in the field from model-based to learning-based approaches, situating the specific technical contributions in a story of field-wide significance that an adjudicator could follow without deep robotics expertise.
The expert letter strategy was also overhauled. The first petition included three expert letters — from the beneficiary's supervisor, a colleague, and a collaborator. The second petition included seven letters: three from professors at leading robotics programs (MIT, Carnegie Mellon, Stanford), two from industry research directors at major technology companies, one from a venture capital partner at a firm that had invested in robotics startups citing the beneficiary's work, and one from the editor of IEEE Robotics and Automation Letters confirming the significance of the journals in which the beneficiary had published. Each letter writer was documented with a one-page credentials summary establishing their recognition in the field, addressing the Kazarian requirement that expert opinion be grounded in recognized expertise.
Common mistake: The first filing's primary error was rushing to file before the record was fully developed. The beneficiary had legitimate extraordinary ability credentials; the petition failed because those credentials were not translated into the evidentiary language the regulatory standard requires under 8 CFR 214.2(o)(3)(iii)(B). The second petition succeeded not by adding credentials the beneficiary did not have — the underlying achievements were largely the same — but by documenting those achievements with the specificity, third-party corroboration, and contextual narrative that the Kazarian step-two analysis demands. The lesson for practitioners is one that bears repeating in February 2025: the quality of the petition, not just the quality of the beneficiary, determines the outcome.
Lessons for Robotics and Engineering O-1A Practitioners
The robotics engineer's case distills into a set of transferable lessons for practitioners advising STEM clients in February 2025. First, the scholarly article criterion requires not just a list of publications but an analysis of each publication venue's standing in the field. For engineering disciplines where conference proceedings are the primary publication format, the petition must proactively address the distinction between proceedings and archival journals and provide metric-based evidence of conference significance. Adjudicators unfamiliar with the field will not supply this context independently.
Second, patent evidence must be presented as impact evidence, not inventory. The number of patents is less important than the measurable significance of those patents: licensing revenue, commercial deployment, forward citations, and industry adoption. A robotics engineer with three highly cited, commercially licensed patents has stronger evidence than one with twenty patents that have never been cited or deployed. The petition should reflect this distinction explicitly.
Third, comparable evidence is available for achievements that do not fit standard criteria and should be used rather than abandoned. Open-source contributions, technical blog posts that have shaped industry practice, conference talks that have been widely cited in subsequent work — these are legitimate comparable evidence items that, when documented with specificity and third-party corroboration, can significantly strengthen the step-two totality-of-evidence showing under 8 CFR 214.2(o)(3)(iv)(B). Practitioners who limit their clients to the enumerated criteria when those criteria are weaker than comparable evidence the client actually possesses are not serving their clients' interests under the full flexibility the regulatory framework affords.