Early-stage researcher on self-evolving LLM trading agents; first author of EvolveTrade (arXiv 2026)
1
papers here
1
reviewed by Evolve
1
first-author
0
corresponding
—
works · OpenAlex
—
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-09-21) 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: the bibliometric record needed to compute research productivity — total publications, venues, and yearly volumes — is not present in the current data. This axis would measure how much the researcher publishes over time.
Impact—
No impact score is available: citation counts and influence metrics are absent from the input data. This axis would measure citations, uptake, and the work's effect on subsequent research.
Focus100/100
The focus score is 100 because the field distribution shows every indexed paper — one of one — in a single field, machine learning (LLM agents for financial trading). That is maximal topical concentration. It should be read cautiously: with only a single 2026 arXiv paper indexed, the perfect score reflects a narrow early record rather than a sustained, multi-paper research program.
Collaboration28/100
Collaboration scores 28, computed from 3 distinct coauthors across the indexed record. The same three names recur — Yumin Choi, Minki Kang, and Sung Ju Hwang — and Kim is first author on the cataloged paper, which is consistent with a small, stable core team rather than a broad collaborative network. As an early-stage pattern, it may broaden as more work is indexed.
Activity—
No activity score is available: only one paper year (2026) exists in the record, so publishing cadence cannot be computed. This axis would measure ongoing research output in recent years.
Reproducibility55/100
Reproducibility scores 55 (reproMean), derived from a single paper. Evolve rates EvolveTrade's reproducibility readiness as medium: reported gains are suggestive — best in 3 of 6 market regimes — but rest on 3 runs over 20-day evaluation windows with no significance tests. Replication therefore looks plausible but unproven until stronger evaluation evidence or released artifacts confirm it.
Research summary
Sehee Kim is an early-stage researcher whose indexed work centers on large language model agents for financial trading. This profile is name-resolved from public bibliographic data linked to ORCID 0000-0002-1815-6957; no institution or country appears in the input data. The Evolve catalog currently holds one paper: EvolveTrade (arXiv, 2026), on which Kim is first author. The paper proposes experience-driven policy refinement in which a trading agent rewrites its own system prompts to adapt to non-stationary markets, combining prompt optimization, tool use, and portfolio management. Evolve's review finds the gains suggestive rather than settled: the self-rewriting approach performs best in 3 of 6 market regimes, but the evidence rests on 3 runs over 20-day windows with no significance tests, and reproducibility readiness is rated medium (reproducibility score 55, from a single paper). Bibliometric signals are narrow but internally consistent: 100% of indexed work falls in one field — machine learning (LLM agents for financial trading) — yielding a focus score of 100, while collaboration scores 28 with 3 distinct coauthors; repeat collaborators include Yumin Choi, Minki Kang, and Sung Ju Hwang. Output, impact, and activity scores are missing because the underlying bibliometric data is not yet available. Read this profile as an early snapshot that should change quickly as more work is indexed.
Collaboration pattern
The indexed record shows a compact collaboration pattern: 3 distinct coauthors, all recurring — Yumin Choi, Minki Kang, and Sung Ju Hwang — and a collaboration score of 28, consistent with a small core team. Kim is first author on the cataloged EvolveTrade paper (not marked corresponding), suggesting a leading hands-on role within this group. The collaborator set is too small to infer community position; expect it to expand if more work is indexed.
Research keywords
Highlights
Experience-driven self-evolution for LLM trading agents
Suggestive gains, thin evidence
Medium reproducibility readiness
Timeline
- 2026First-authored EvolveTrade appears on arXiv, proposing experience-driven policy refinement for self-evolving LLM trading agents.
- 2026Evolve review records suggestive gains (best in 3 of 6 regimes) with thin evidence: 3 runs, 20-day windows, no significance tests.
- 2026Reproducibility readiness is rated medium (score 55); the researcher's Evolve profile opens as an early-stage, single-field record.
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 scores are computed from OpenAlex-based data only; no other bibliographic sources are reflected.
- The catalog contains a single paper (2026 arXiv), so output, impact, and activity dimensions are missing and every present score rests on one record.
- Institution and country are absent from the input data; the profile is name-resolved from public bibliographic data linked to ORCID 0000-0002-1815-6957.
- Sehee Kim is a relatively common Korean name; identity continuity beyond the ORCID-linked record is not verified here.