CV
Contact Information
| Name | Zheqing Zhang |
| Professional Title | PhD Student, AI Researcher, Alpha Digger |
| zhangzheqing@ruc.edu.cn | |
| Location | Beijing, China |
Professional Summary
PhD student at Gaoling School of AI, Renmin University of China. Research focuses on LLM Agents, Deep Research, and AI4Finance.
Experience
-
2025.02 - 2025.06 Beijing, China
LLM Algorithm / Backend Dev Intern
AIGility Cloud Innovation 智悦云创 (AI Startup)
Core member of the “Resume Optimizer” product line, responsible for backend architecture design and Agent orchestration.
- Proposed TiMem temporal-hierarchical memory framework, solving agent memory decay in long conversations. Accepted by ACL Findings 2026.
- TiMem framework successfully deployed as the company’s main product line.
Education
-
2026 - Beijing, China
-
2022 - 2026 Beijing, China
Bachelor
Renmin University of China · Gaoling School of AI
Artificial Intelligence
- C++: 91 AI Systems: 93 Recommendation Systems: 93 Introduction to Information Retrieval: 91
- Second-Class Scholarship for Social Work and Volunteer Service
- Outstanding Graduate
- Outstanding Student Cadre
- Outstanding League Cadre
- {“TOEFL”=>99}
Publications
-
2025 Investigating and Extending Homans' Social Exchange Theory with Large Language Model based Agents
ACL 2025 (Oral)
Proposed a computational sociology paradigm using LLM Agents, successfully verified Social Exchange Theory in agent groups and extended traditional theory.
-
2026 TiMem: Temporal-Hierarchical Memory Consolidation for Long-Horizon Conversational Agents
ACL Findings 2026
Proposed TiMem, a temporal-hierarchical memory framework that consolidates fragmented conversations into structured persona profiles, now deployed as a main product line.
Skills
Languages
Interests
Projects
-
BullClaw
Multi-language Financial Report Generation System based on LLM Multi-Agents. Built stateful multi-agent workflow using LangGraph, supporting 8 global exchanges in 4 languages.
- OpenFinArena Ranking: Overall #3, US Market #1, China #2
- Outperformed DeepSeek-v3.2 and other baselines
-
Stock-Meme-Hunter
Hype Logic Mining based on Hybrid Retrieval and LLM Reranking. Proposed pinyin + concept dual inverted index for A-share metaphorical hype.
- Designed 4-layer funnel model balancing coverage and cost
Awards
-
2026 Outstanding Graduate
Renmin University of China
-
2025 Outstanding Student Cadre
Renmin University of China
-
2025 Outstanding League Cadre
Renmin University of China
-
2025 Second-Class Scholarship for Social Work and Volunteer Service
Renmin University of China