Early-career UBC researcher on LLM agents for algorithmic trading and inference-as-search, with a focused NLP record
2
papers here
1
reviewed by Evolve
0
first-author
0
corresponding
—
works · OpenAlex
1
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-10-17) 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—
Output metrics are not yet available: this axis would measure total publication volume, venue quality, and first- or corresponding-author share. The bibliometric snapshot provides no reliable output score, so no estimate is offered here.
Impact—
Impact cannot be scored yet: this axis would track citation counts, h-index momentum, and uptake of methods by other groups. The snapshot shows only h-index 1 and i10-index 1 — too sparse for a meaningful impact assessment.
Focus100/100
Focus is the strongest axis at 100. OpenAlex topic assignments concentrate tightly: Topic Modeling and Natural Language Processing Techniques lead with 5 works each, Advanced Text Analysis adds 2, and the reviewed catalog field is machine learning (LLM agents for algorithmic trading). Two single-count topics — electrocatalysts for energy conversion and financial markets — barely dilute an otherwise sharply defined NLP-and-agents identity.
Collaboration58/100
The collaboration score of 58 reflects a moderately networked early career: 17 distinct co-authors across 6 works in the last three years. Recurring partners include Jiaqi Wei, Xiang Zhang, and Wenxuan Huang (each appearing repeatedly in the co-author list), alongside Nuoyue Xu, Jiang Liu, Yuejin Yang, Sheng Xu, Xiang Zhuang, and Zhangyang Gao — a pattern consistent with small team-based arXiv projects rather than large consortia.
Activity57/100
Activity scores 57, driven by a recent uptick: the indexed record shows 1 work in 2020, 4 in 2025, and 2 in 2026 to date, with 6 works in the trailing three years. Output is concentrated in the 2025–2026 window, suggesting an accelerating but still low-volume publishing rhythm for an early-stage profile.
Reproducibility55/100
Reproducibility scores 55 from a single assessed paper (reproPapers 1, mean readiness 55): the episodic-retrieval trading agent, labeled medium. Evolve's review finds its claimed +3.5–22pp accuracy gains plausible but statistically untested, with thin baselines and a mislabeled headline effect size — factors that cap confidence in independent replication.
Research summary
Juntai Cao (University of British Columbia, Canada) is an early-stage researcher working at the intersection of large language model agents and quantitative finance, with a bibliometric record that also shows roots in topic modeling and natural language processing. OpenAlex topic data assigns 5 works each to Topic Modeling and Natural Language Processing Techniques, 2 to Advanced Text Analysis, and single works to electrocatalysts and financial-markets topics, yielding a maximum focus score of 100. Output is recent and accelerating: 1 indexed work in 2020, 4 in 2025, and 2 in 2026 to date, with 6 works in the trailing three years and 17 distinct co-authors; the citation footprint remains minimal (h-index 1, i10-index 1), so output and impact dimensions are marked missing. The Evolve-reviewed catalog contains two 2026 arXiv papers, both with Cao in a middle author position: an episodic-retrieval memory module for LLM trading agents claiming +3.5–22pp accuracy gains — judged plausible but statistically untested, with thin baselines and a mislabeled headline effect size (reproducibility readiness: medium) — and a systematic treatment of inference as tree search in LLMs that awaits review. Overall this is a focused, fast-moving but early profile whose reproducibility signal (55 from one assessed paper) rests on limited evidence.
Collaboration pattern
The record lists 17 distinct co-authors over the last three years (6 works), a moderately wide but small-team network. Jiaqi Wei, Xiang Zhang, and Wenxuan Huang each recur in the frequent-coauthor list, suggesting stable working pairs, while Nuoyue Xu, Jiang Liu, Yuejin Yang, Sheng Xu, Xiang Zhuang, and Zhangyang Gao appear alongside. Both reviewed papers place Cao in a middle author position without corresponding-author status, consistent with a contributor role within arXiv-stage project teams.
Research keywords
Highlights
Episodic-retrieval memory for LLM trading agents claims +3.5–22pp gains
A systematic view of inference as tree search in large language models
Cao's second 2026 arXiv catalog entry (middle authorship) frames LLM inference through the lens of tree search. Evolve has not yet published a verdict, keywords, or a reproducibility rating for this paper, so its contribution remains unrated in this profile.
Timeline
- 2020First work appears in the indexed record (1 publication).
- 2025Most productive indexed year, with 4 publications recorded.
- 20262 publications indexed so far, including both Evolve-reviewed arXiv papers on episodic-memory trading agents and inference as tree search.
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.
- Identity is name-resolved from OpenAlex bibliographic records; no ORCID is linked, so same-name works may be merged or missing.
- All bibliometrics (h-index, i10-index, per-year counts, topics, co-author counts) come from the OpenAlex snapshot only.
- Output and impact dimensions lack sufficient underlying data and are reported as missing, not estimated.
- Only one catalog paper carries an Evolve verdict and reproducibility score; the tree-search paper is unrated, so quality signals rest on a single data point.