First-Year MPhil Student

Jiaqi Yuan

I'm interested in self-improving AI agents that learn, reason, and evolve through interaction with data and the world.

Portrait of Jiaqi Yuan
01

About

I am currently an MPhil student at The Hong Kong University of Science and Technology (Guangzhou), advised by Professor Zhijiang Guo. Before that, I graduated from the University of Electronic Science and Technology of China (UESTC).

My research focuses on building autonomous and continually improving AI systems, with particular interests in agentic intelligence, data agents, and foundation model training.

Agentic AI Self-Improving Data Agents Foundation Model Training Techniques

A little more about me: my MBTI is INFP.

02

Experience

Research Intern

Ling Foundation Model Pretraining · Ant Group

Worked on pretraining techniques for the Ling foundation model.

03

Selected Papers

Selected leading works.

2026
Overview of the AgenticGEO framework
First author

AgenticGEO: A Self-Evolving Agentic System for Generative Engine Optimization

J. Yuan, J. Wang, Z. Wang, Q. Sun, R. Wang, J. Li · 2026

A self-evolving agentic framework for Generative Engine Optimization that maintains a diverse strategy pool and co-evolves a lightweight surrogate critic to reduce costly black-box engine feedback.

PDF ↗ Citations 3 GitHub ★ 40
2025
Overview of the DyG-RAG framework
First student author

DyG-RAG: Dynamic Graph Retrieval-Augmented Generation with Event-Centric Reasoning

Q. Sun, J. Yuan, S. He, X. Guan, H. Yuan, X. Fu, J. Li, P. S. Yu · 2025

The first temporal GraphRAG framework to model temporal knowledge as a dynamic event graph, using Dynamic Event Units for time-aware retrieval and Time Chain-of-Thought reasoning.

PDF ↗ Citations 10 GitHub ★ 59

Let's talk

Open to ideas and collaboration.

If you are interested in my research or have a potential collaboration in mind, feel free to get in touch.