I'm an undergraduate at UC San Diego studying Data Science and Statistics.
My research centers on a fundamental question: how can AI systems keep learning after deployment? I build self-evolving agents capable of test-time continual learning—systems that accumulate concepts, abstract from experience, and know when to trust themselves.
My work spans memory-augmented LLM agents, abstract reasoning, AI for scientific discovery, and uncertainty quantification.
Latest Research
View All →Research Assistant @ Q-Lab
Investigating memory systems for compositional reasoning on ARC-AGI. Designed feature extraction methodology for Program Synthesis memory format.
Research Assistant @ Wang Lab
Architected a heterogeneous Mixture-of-Experts (MoE) system for chemical reaction prediction. Implemented GNN encoders with shortest-path positional encodings.
Selected Publication
View All →ArcMemo: Abstract Reasoning Composition with Lifelong LLM Memory
Matthew Ho, Chen Si, Zhaoxiang Feng, Fangxu Yu, Yichi Yang, Zhijian Liu, Zhiting Hu, Lianhui Qin.