Vipin Chaudhary
Case Western Reserve University · United States
Case Western Reserve University researcher in LLM reliability, evaluation, and AI governance
4
papers here
4
reviewed by Evolve
0
first-author
0
corresponding
—
works · OpenAlex
1
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-10-16) 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—
The output axis would measure total publication volume and long-run productivity, but the available record supplies only a partial recent-3-year sample rather than verified full-output counts, so this axis is left unscored rather than estimated.
Impact—
The impact axis would measure citation-based influence such as citation counts and h-index trajectory. OpenAlex currently shows an h-index of 1 and an i10-index of 0 — too thin, and possibly distorted by name disambiguation, to support a reliable score.
Focus0/100
The focus score of 0 reflects an evenly spread OpenAlex field distribution: LLM-based multi-agent systems, AI governance and security (LLM text watermarking), formal theorem proving / machine learning for mathematics, and neurosymbolic AI each hold 0.25. Topic labels (topic modeling 7, NLP techniques 6, multimodal ML 6) confirm breadth across themes rather than concentration in one area.
Collaboration59/100
Collaboration scores 59, grounded in 18 distinct co-authors across 19 works in the last three years. A recurring cluster of twelve frequent collaborators — including Debargha Ganguly, Vikash Singh, Yu-Neng Chuang, Zhengchen Li, and Erman Ayday — anchors the output; on the four reviewed papers this researcher appears in middle or last author positions and is corresponding author on none.
Activity87/100
Activity is the strongest axis at 87, driven by 19 works over the last three years and a countsByYear profile of 1 (2023), 1 (2025), and 18 (2026) — a sharp 2026 surge that includes all four Evolve-reviewed arXiv preprints. Topic counts (topic modeling 7, NLP techniques 6, multimodal ML 6) reflect this recent burst.
Reproducibility66/100
Reproducibility scores 66, the mean readiness across the 4 Evolve-reviewed papers. The multi-agent efficiency re-evaluation is rated high readiness, while the autoformalization reward-model, watermarking, and StochBench papers are rated medium — consistent with an evaluation-heavy portfolio whose verdicts stress controlled comparisons and candidly reported limitations.
Research summary
Vipin Chaudhary is a researcher at Case Western Reserve University (US) whose Evolve catalog centers on the reliability, evaluation, and governance of large language model systems. The OpenAlex-resolved record shows 19 works in the last three years, with a sharp concentration in 2026 (18 works versus 1 each in 2023 and 2025), and an evenly weighted four-field portfolio: LLM-based multi-agent systems, AI governance and security around LLM text watermarking, formal theorem proving / machine learning for mathematics, and neurosymbolic AI (0.25 each). Evolve has reviewed four 2026 arXiv preprints: generative reward models that detect solver-invisible unfaithful SMT encodings at 0.961 AUROC; a policy analysis arguing that non-verifiability, not watermarking itself, is the real EU AI Act gap; StochBench, a 450-target Lean 4 benchmark for graduate stochastic processes where an Opus 4.8 agent reaches 34.9%; and a controlled re-evaluation finding that reported gains from five multi-agent pruning methods are setup-dependent. The work is evaluation-heavy and benchmark-driven, with a mean reproducibility readiness of 66 across the four papers. Citation aggregates remain early-stage — h-index 1 and i10-index 0 on OpenAlex — so impact cannot yet be scored, and the record may be affected by name disambiguation.
Collaboration pattern
Collaboration scores 59: 18 distinct co-authors across 19 works in the last three years, with a stable recurring cluster of 12 frequent collaborators including Debargha Ganguly, Vikash Singh, Yu-Neng Chuang, Zhengchen Li, Jiamu Zhang, Hanjie Chen, and Erman Ayday. On the four Evolve-reviewed papers this researcher sits in middle or last author positions and is corresponding author on none, consistent with a contributing specialist within larger LLM-evaluation teams.
Research keywords
Highlights
Generative reward models catch solver-invisible unfaithful formalizations
Arguing that verification, not watermarking, is the EU AI Act gap
StochBench: a 450-target Lean 4 benchmark for stochastic processes
Multi-agent efficiency gains shown to be setup-dependent
Timeline
- 2023OpenAlex records 1 indexed work in 2023, the earliest year in the available countsByYear data.
- 2025OpenAlex records 1 indexed work in 2025, preceding the sharp rise in output the following year.
- 2026A marked surge: 18 indexed works in 2026, including all four Evolve-reviewed arXiv preprints — the GenV+HN autoformalization reward model, the EU AI Act watermarking analysis, StochBench for Lean 4, and the re-evaluation of multi-agent efficiency methods.
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 indicators come from OpenAlex only; the reported h-index of 1 and i10-index of 0 appear inconsistent with 19 recent-3-year works, so the identity record may be split or conflated by name disambiguation.
- No ORCID was provided; this profile is name-resolved from public bibliographic data, and citation data is too thin to score the output and impact axes.
- All four reviewed papers are 2026 arXiv preprints; peer-review status and published venues are not reflected in this snapshot.
- The OpenAlex field distribution is evenly split across four fields (0.25 each), so the three listed fields are a subset of a broader, evenly weighted portfolio.