Dongming Jiang
The University of Texas at Dallas · United States
UT Dallas AI-systems researcher making LLM agents and inference cheaper via agentic memory and cross-model KV reuse
2
papers here
1
reviewed by Evolve
1
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—
Not yet scored: verified output data beyond the two reviewed arXiv papers is unavailable; this axis would measure total publication volume, venue distribution and productivity composition once a complete publication record is assembled.
Impact—
Not yet scored: citation impact cannot be meaningfully assessed — the snapshot shows h-index 1 and i10-index 0; this axis would measure citation uptake and scholarly influence as the publication record matures.
Focus100/100
Scores 100/100. The OpenAlex topic profile is tightly concentrated: Information Retrieval and Search Behavior (9 works), Multimodal Machine Learning Applications (8) and Topic Modeling (7) lead, ahead of Advanced Neural Network Applications (4) and Recommender Systems (4); the catalog field distribution lists a single field, AI systems (LLM agents and memory architectures). Both reviewed papers converge on efficient LLM memory and context reuse, reinforcing a sharply defined agenda.
Collaboration36/100
Scores 36/100. The co-author network is small: the snapshot records only 5 distinct co-authors, with Yi Li, Bingzhe Li and Yi Zhao recurring as frequent collaborators. This compact, stable core-team pattern — with first authorship on Jev-Mem and a middle position on the context-reuse layer paper — is typical of an early-stage researcher rather than a broad multi-institution network.
Activity83/100
Scores 83/100. Output cadence is high and accelerating: OpenAlex counts-by-year shows 3 works in 2025 rising to 13 in 2026, with 16 works within the recent three-year window. Both reviewed arXiv papers appeared in 2026, including a first-authored system paper, consistent with a fast-moving, high-tempo production pace.
Reproducibility55/100
Scores 55/100, computed from a single assessed paper (reproMean 55 across 1 paper). Evolve rates Jev-Mem medium — a promising efficiency design evaluated on one benchmark without ablations — while the cross-model KV-sharing paper is low readiness, its translator structure unspecified. With one data point, treat this score as provisional.
Research summary
Dongming Jiang is an early-stage researcher at The University of Texas at Dallas working at the intersection of LLM systems and intelligent information access. The bibliometric snapshot (OpenAlex identity) records 16 works in the recent three-year window, with counts-by-year rising from 3 works in 2025 to 13 in 2026 — an accelerating cadence. The topic profile is highly concentrated: Information Retrieval and Search Behavior (9 works), Multimodal Machine Learning Applications (8), Topic Modeling (7), Advanced Neural Network Applications (4), and Recommender Systems (4). Evolve has reviewed two 2026 arXiv papers. The first-authored Jev-Mem proposes System-One-controlled agentic memory; Evolve's review notes claimed gains of 0.777 on LoCoMo (+11% relative) with 6.6x faster memory builds, while flagging a single benchmark and missing ablations (reproducibility: medium). The co-authored A Universal Context-Reuse Layer for Cross-Model KV Sharing reports up to 67% reduction in target prefill cost across 3 model pairs at near-baseline quality, with the translator structure unspecified (reproducibility: low). Formal citation indicators remain nascent (h-index 1, i10-index 0), so output and impact are unscored. Collaboration is compact — 5 distinct co-authors, with Yi Li, Bingzhe Li and Yi Zhao recurring. The profile coheres around one theme: making LLM agents and inference cheaper through memory and context reuse.
Collaboration pattern
The snapshot records 5 distinct co-authors, a deliberately small pool; Yi Li, Bingzhe Li and Yi Zhao recur as frequent collaborators, suggesting a stable core team. Both reviewed 2026 arXiv papers are multi-authored, with Jiang first author on Jev-Mem and in a middle position on the context-reuse layer. No cross-institution or senior-collaborator patterns are visible in the provided data.
Research keywords
Highlights
Jev-Mem: System-One-controlled agentic memory for efficient AI agents
A universal context-reuse layer for cross-model KV sharing
Timeline
- 2025Bibliometric record shows 3 works in 2025, the earliest year in the OpenAlex counts-by-year snapshot.
- 2026Output rises to 13 works in 2026 per the OpenAlex counts-by-year, with 16 works across the recent three-year window.
- 2026First-authors Jev-Mem, a System-One-controlled agentic memory paper on arXiv, reviewed by Evolve with medium reproducibility readiness.
- 2026Co-authors A Universal Context-Reuse Layer for Cross-Model KV Sharing on arXiv, reviewed by Evolve with low reproducibility readiness.
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 (identity: openalex); no ORCID is linked, and Dongming Jiang is a common name, so this profile is name-resolved from public bibliographic data and may not capture every work by this name.
- Output and impact dimensions are unscored: citation data is limited to h-index 1 and i10-index 0, and the full publication record beyond the two reviewed arXiv papers is not available.
- The reproducibility score (55) rests on a single assessed paper; treat it as provisional.
- Only 2 papers from the Evolve catalog have been reviewed; topic and counts-by-year data reflect the OpenAlex snapshot and may include works not yet reviewed.