Jie Cao — UNC Chapel Hill researcher working across robotics education and multimodal AI agents
2
papers here
2
reviewed by Evolve
0
first-author
0
corresponding
—
works · OpenAlex
—
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-10-08) 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 output score is available yet. This axis would measure publication volume and productivity over time — total papers, venues, and authorship roles — but the current bibliometric snapshot contains too few indexed records to compute it, so no figure is reported.
Impact—
No impact score is available yet. This axis would measure citation-based influence — citations, h-index-style aggregates, and uptake across fields — but no citation data accompanies the current snapshot, so no estimate is made.
Focus0/100
Focus currently scores 0 because the field distribution is an exact 0.5/0.5 split: robotics and engineering education on one side, machine learning with multimodal AI agents on the other. With only two cataloged papers, this reads as a deliberate dual-track portfolio rather than an established specialty, and a single dominant field may still emerge as output grows.
Collaboration39/100
The collaboration score is 39, built on a small pool of six distinct co-authors across the two cataloged papers. Recurring partners include Yuanzhe Dong, Shuman Wang, Hexiong Yang, Mingrui Chen, and Ran He. Cao holds a middle, non-corresponding author position on both papers, suggesting contributor-level involvement within modest-sized teams rather than leading large consortia.
Activity—
No activity score is available yet. This axis would measure publishing cadence and recent momentum — papers per year and year-over-year consistency — but the cataloged record spans only 2026, which is too short a window to score.
Reproducibility38/100
Reproducibility averages 38 across the two papers Evolve has assessed (reproMean 38 over reproPapers 2). The spread matters: the high-school humanoid RL curriculum is rated low readiness, consistent with its verdict that effectiveness remains unevaluated, while VLM-in-Sandbox reaches medium readiness with a training-free, rigorously evaluated framework. With only two data points, the score is indicative rather than definitive.
Research summary
Jie Cao is affiliated with the University of North Carolina at Chapel Hill, and the reviewed record depicts an early-stage researcher working across two distinct areas: robotics and engineering education, and machine learning centered on multimodal AI agents. The Evolve catalog contains two 2026 arXiv preprints, with Cao in a middle author position and not the corresponding author on either. The first is an expert-validated curriculum for teaching reinforcement learning and humanoid robotics to high-school students on a low-cost open platform; review judged it honest and credible while noting that classroom effectiveness remains unevaluated. The second, VLM-in-Sandbox, introduces a training-free visual workspace for sandboxed vision-language-model agents; its evaluation was rated rigorous and honest, though the promote mechanism's accuracy gain is statistically inconclusive. The bibliometric radar mirrors this dual-track profile: a focus score of 0 from an even 0.5/0.5 field split, a collaboration score of 39 resting on six distinct co-authors, and a reproducibility score of 38 averaged over two assessed papers — low readiness for the curriculum, medium for the sandbox framework. Output, impact, and activity axes lack sufficient data. The profile should be read as an early snapshot built on a small, preprint-only catalog.
Collaboration pattern
Collaboration is modest in scale: the two cataloged papers share a pool of six distinct co-authors, giving a collaboration score of 39. Recurring names include Yuanzhe Dong, Shuman Wang, Hexiong Yang, Mingrui Chen, and Ran He. Cao sits in a middle author position on both papers without corresponding-author status, consistent with contributor-level roles in small project teams rather than leading large consortia.
Research keywords
Highlights
VLM-in-Sandbox: training-free visual workspaces for sandboxed VLM agents
Expert-validated high-school curriculum for RL and humanoid robotics
Timeline
- 2026Published an expert-validated curriculum teaching reinforcement learning and humanoid robotics to high-school students on a low-cost open platform (arXiv).
- 2026Released VLM-in-Sandbox, a training-free visual workspace for sandboxed vision-language-model agents (arXiv).
- 2026Cataloged output spans two fields in the same year — robotics and engineering education, and machine learning (multimodal AI agents) — each contributing half of the field distribution.
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 radar scores are computed from a limited public bibliometric snapshot; only collaboration, focus, and reproducibility could be scored.
- The reviewed catalog contains only two 2026 arXiv preprints; no citation counts, venue history, h-index, or career facts are available in this data.
- "Jie Cao" is a common name; this profile is name-resolved from public bibliographic data and may merge or omit same-name records.
- Reproducibility readiness (mean 38) is based on only two assessed papers, so the score is indicative rather than definitive.