M.S. student in Automation at Shanghai Jiao Tong University, Class of 2029. I focus on building reliable LLM applications, multi-agent systems, long-term memory, and GraphRAG.
- Former LLM Application R&D Intern at Meituan Infrastructure R&D Platform
- Owned the 0-to-1 delivery of a multi-agent long-term memory and personalized context system
- Worked on heterogeneous agent integration, memory extraction, user profiling, retrieval, and context injection
- Built GraphRAG multi-agent workflows with Neo4j and LangGraph
| Agent Memory Extraction, storage, retrieval, and reflection |
Personalization User profiling and context engineering |
GraphRAG Knowledge graphs and multi-hop retrieval |
Agent Systems Orchestration, evaluation, and reliability |
|
Python research tooling for evaluating SSO referral-flow abuse resistance, quota behavior, token lifecycle, and activation telemetry in controlled environments. 
|
Dockerized Keycloak and SAML SSO deployment with automated workspace configuration and user provisioning. 
|
Machine-learning weather prediction, data processing, and interactive visualization with Python and Streamlit. 
|
I am currently exploring memory evaluation, personalized retrieval, context engineering, and reliable orchestration for production agent systems.
