Kairan Jin

M.S. in Management Science & Engineering · USTC · AI Algorithms
I work at the intersection of operations research, evolutionary computation and LLM-based multi-agent systems. My research focuses on designing self-improving heuristics for large-scale combinatorial optimization (retired-battery repurposing, urban vehicle platooning), and on building agentic frameworks that turn domain knowledge into reproducible research artifacts.
Kairan Jin

Selected Publications 3 items

CIE
under review

Capacity-Constrained Battery Repurposing Research

Jin, K. et al.
Highlight Auto-evolved heuristics via AlphaEvolve, reaching a <0.52% optimality gap versus exact MILP on stochastic multi-attribute battery reassembly problems.
Formulates the retired-battery repurposing problem as a multi-attribute stochastic MDP; capacity-feasible allocation / neighborhood search / feasibility filtering are modularized as evolvable operators inside a combinatorial heuristic search space.
Computers & Industrial Engineering · JCR Q1 · First author · Under review
🔗 SSRN Preprint
MDPHeuristics AlphaEvolveCombinatorial Optimization Circular Economy
ITSC 2023
ieee conference

Distributed Model Predictive Control Considering Delay-Compensating Strategy for Urban Vehicle Platooning in A Mixed Autonomy Traffic

Liu, M., Jin, K., Hao, R.
A distributed MPC scheme with explicit communication-delay compensation for connected & autonomous vehicle platoons operating in mixed-autonomy urban corridors.
IEEE ITSC 2023 · Second author (advisor as first) · DOI: 10.1109/ITSC57777.2023.10422427
🔗 IEEE Xplore
MPCPlatooning Mixed AutonomyDelay Compensation
TSE&I
CSSCI / 北大核心

城市道路智能网联车辆轨迹鲁棒控制方法

刘美岐, 金楷然, 李雅澜, 等
Proposes a robust trajectory-control method for connected & autonomous vehicles on urban arterials, bounding worst-case tracking error under model mismatch and sensing noise.
交通运输系统工程与信息 · 2024, 24(4): 31–40 · 北大核心 · Second author (advisor as first)
🔗 Journal
Robust ControlCAV Trajectory Planning

Research Projects 4 items

AlphaEvolve × Battery Repurposing

2024 – 2025 · Master's thesis

Auto-evolved heuristics for capacity-constrained retired-battery reassembly. Combines a multi-attribute stochastic MDP with an AlphaEvolve outer loop that mutates heuristic operators (allocation, neighborhood search, feasibility filtering). Reaches <0.52% gap vs. the exact solver.

AlphaEvolveMDPHeuristics

Urban Platoon Cooperative Control

2022 – 2024 · Undergraduate research

Distributed cascaded coordination between traffic-signal control and CAV platoon trajectory planning in mixed-autonomy urban networks. Produced one IEEE ITSC paper and one CSSCI-core journal article.

MPCSignal ControlRobust

ContentBuddy · Multi-Agent Writing

2025 · Tencent AI Hackathon (4/117)

A 7-Agent architecture for creator-oriented long-form writing: hot-topic tracking (Zhihu / Toutiao / Weibo), 4-channel search (Google / Zhihu / BoCha / Tavily), adversarial critic review and multi-platform style adaptation. −70% topic-research time; per-article cost under ¥0.5.

LangGraphMulti-AgentSSE

Agent SFT · Function Calling

2025 · Multi-turn tool-use fine-tuning

Qwen3-0.6B + LoRA SFT with 3,120 seed → 7,612 turn-split samples, turn-by-turn loss weighting to fix tool-call over-triggering. Argument JSON validity 58% → 96%; end-to-end success 27% → 94%.

Qwen3LoRASFT

SemiResearch (DeepResearch)

2026.01 – 2026.04 · CXMT R&D

A domain-specialized multi-agent system for semiconductor research reports. Six LangGraph agents (plan / retrieve / analyze / chart / write / audit) with hybrid retrieval, credibility scoring and expert-critic feedback loops.

LangGraphRAGMulti-Agent

Education & Awards

Education

M.S. · University of Science and Technology of China
Management Science & Engineering
2024.09 – 2027.06 · 合肥
B.Mgmt · Dalian Maritime University
Supply Chain Management · Rank 1 / class
2020.09 – 2024.06 · 大连

Selected Awards

🏆 USTC 研究生一等奖学金 🏆 国家奖学金 ×2 🏆 辽宁省优秀毕业生 🏆 校三好学生 🥇 高校商业挑战赛 · 国家一等奖 🥈 MCM/ICM · Meritorious (二等奖) 🥈 服务外包大赛 · 国家二等奖

Reviewer / Service

Currently focusing on multi-agent evaluation methodology and heuristic-generation benchmarks. Open to reviewing / collaboration on OR × LLM topics — feel free to reach out.