Hi there! I am a third-year PhD student in Computer Science at Cornell University, advised by Claire Cardie and Sainyam Galhotra.

Previously, I obtained my B.S. and M.S. degrees from Jilin University and Sun Yat-sen University respectively, both in computer science. I worked with Professor Hui Xiong from Hong Kong University of Science and Technology on weakly supervised machine learning for graph data during my master's studies. I interned as an applied scientist at AWS during 2026 summer, working on agentic Text-to-SQL.

My recent research focuses on LLM reasoning and agentic systems. I am particularly interested in how language models learn and retain reasoning capabilities, and how AI agents retrieve information, resolve uncertainty, and interact with complex data environments. My earlier work spans machine learning theory and graph representation learning. Representative research directions include:

• LLM Reasoning & Post-Training, where I study how reinforcement learning affects reasoning diversity and develop methods to preserve diverse, correct reasoning behaviors while improving performance. [TAR]
• Agentic Data Management Systems, where I develop LLM agents that discover relevant information across complex data lakes and resolve underspecified user requests through interactive clarification. [Octopus, PlanPool]
• Machine Learning Theory, where I studied query-efficient approximation algorithms for metric clustering and graph optimization using noisy, weak comparison oracles. [COLT'25]
• Machine Learning on Graphs, where I developed methods for learning effective and robust representations of graph-structured data, including imbalanced node classification, graph contrastive learning, and heterogeneous graph modeling. [GraphSHA, HomoGCL, AEGCL, StarGAT]