Google researcher working across LLM-assisted RTL verification, hardware security, and retrieval-augmented generation
2
papers here
2
reviewed by Evolve
0
first-author
0
corresponding
—
works · OpenAlex
—
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-10-15) and cover the author's entire public record, not just the papers listed here. Reproducibility signal comes from Evolve's own reviews.
Research profile
Output—
No publication-volume data is available for this axis yet. It would measure the size and cadence of research output — total papers, preprints versus peer-reviewed venues, and yearly counts. Only two reviewed arXiv papers are in scope.
Impact—
No citation data is available for this axis yet. It would measure citations, citing works, and uptake in hardware-security and IR communities; both reviewed papers are 2026 preprints, too recent for impact signals.
Focus0/100
The observed record is evenly split rather than concentrated: the two reviewed 2026 papers divide 0.5/0.5 between hardware security and AI for EDA (RTL verification) and retrieval-augmented generation for IR/NLP, producing a focus score of 0/100. This reads as an early-stage, two-track exploratory profile — auditing LLM-generated verification artifacts on one track, source-risk triage for RAG on the other — rather than an established single-field specialization.
Collaboration44/100
Collaboration scores 44/100 on a small-team pattern: 8 distinct coauthors appear across the reviewed record, with recurring partners Chuhong Xu, Kainan Zhou and Lu Yi, plus Hang Xiao. On both catalog papers Gangzhen Qian sits in a middle author position without corresponding-author status, consistent with contributing within a modest, stable core group rather than leading large or independent efforts.
Activity—
No year-by-year publication data is available for this axis yet. It would measure research tempo — papers per year, gaps, and sustained engagement across the career. The observed record covers only 2026, so activity cannot be scored.
Reproducibility38/100
Reproducibility readiness averages 38/100 across the 2 reviewed papers. The preregistered RTL-verification audit is rated medium — its fail-closed benchmark and baseline protocol remain usable even though the AI runs yielded no valid analysis — while the RAG source-risk triage paper is rated low, resting on a single-coded pilot plus a synthetic scenario. Artifacts exist, but independent replication demands are high.
Research summary
Gangzhen Qian is a researcher at Google whose reviewed work spans two distinct technical tracks: hardware security and AI for EDA, focused on LLM-assisted RTL verification, and retrieval-augmented generation (RAG) for information retrieval and NLP. The Evolve catalog holds two 2026 arXiv papers, in both of which Qian appears as a middle author without corresponding-author status. The first, 'Trust, but Validate the Instrument: Auditing AI-Generated RTL Verification Plans on Authored Security-Regression Proxies', is a preregistered audit of LLM-generated verification plans. Its AI runs yielded no valid analysis, so it reports no model-performance evidence, but it is valued as a fail-closed benchmark and baseline built on authored security-regression proxies, with medium reproducibility readiness. The second, 'Query-Aware Source-Risk Triage for Retrieval-Augmented Generation', proposes auditable, query-aware source-risk scores combining query clustering with rank-discounted aggregation; its evidence base is a single-coded pilot plus a synthetic scenario, and its reproducibility readiness is rated low. Bibliometrically this is an early-stage, two-track profile: field distribution splits evenly at 0.5/0.5 between the two fields, yielding a focus score of 0/100; collaboration scores 44/100 across 8 distinct coauthors; mean reproducibility readiness is 38/100 over 2 papers. Output, impact and activity lack sufficient underlying data to score.
Collaboration pattern
The reviewed record shows a compact collaboration pattern: 8 distinct coauthors, with a recurring core of Chuhong Xu, Kainan Zhou and Lu Yi, plus Hang Xiao. Gangzhen Qian sits in a middle author position on both 2026 arXiv papers and is not listed as corresponding author on either, suggesting a contributor role within a small, stable working group spanning hardware-security verification and RAG research rather than a solo or large-consortium mode.
Research keywords
Highlights
A fail-closed audit of LLM-generated RTL verification plans
Query-aware source-risk triage makes RAG sources auditable
Timeline
- 2026Publishes 'Trust, but Validate the Instrument: Auditing AI-Generated RTL Verification Plans on Authored Security-Regression Proxies' on arXiv, a preregistered audit in hardware security and AI for EDA (middle author).
- 2026Publishes 'Query-Aware Source-Risk Triage for Retrieval-Augmented Generation' on arXiv, proposing auditable source-risk scores for RAG (middle author).
- 2026Both reviewed outputs land in the same year across two distinct fields — RTL hardware-security verification and retrieval-augmented generation — anchoring an early-stage, two-track profile.
Papers in the Evolve catalog
Frequent co-authors
About this profile
This profile is AI-structured by Evolve.Pub from public bibliographic data (OpenAlex) and Evolve's own paper reviews. It is not written or endorsed by the researcher unless the profile is claimed. Radar dimensions are computed from real bibliometrics with fixed formulas — raw figures are shown alongside; missing dimensions mean insufficient data, never zero. Errors? Claim the profile to correct or hide it.
- Bibliometric signals are drawn from OpenAlex-derived public bibliographic data only; no citation counts or venue-quality metrics were in scope.
- The output, impact and activity axes lack underlying data (no publication-volume or citation records), so no scores are reported for them.
- The evidence base is two Evolve-reviewed 2026 arXiv preprints; no peer-reviewed venue papers are included in this sample.
- This profile is name-resolved from public bibliographic data via the provided ORCID; 'Gangzhen Qian' is not a unique name, and the affiliation (Google, United States) is as reported in the input data.