I'm a fully-funded PhD candidate at the School of Artificial Intelligence, CUHK-Shenzhen, advised by Prof. Feng Yin and Prof. Tsung-Hui Chang. Before that, I did my BSc in Statistics (first-class honors) at CUHK-Shenzhen. Along the way I've interned at ByteDance, Huawei, and Dria.
My research is on adaptive inductive biases for data-efficient optimization and discovery, assumptions that can be learned from limited data, revised during search, and proposed by open-ended search.
Lately, the questions on my mind are
- How can models adapt their assumptions to the task when data and evaluations are limited?
- How can AI agents discover new algorithms and models that we can verify?
- How do we evaluate AI research systems on real problems and improve them from feedback?
More on my research page.
| Project | What it does |
|---|---|
| Kernaut | Coding agents that discover new kernels as programs |
| CAKE | LLMs that evolve the kernel during Bayesian optimization (NeurIPS 2025) |
| PlugBO | A modular framework that lets an agent reconfigure Bayesian optimization on the fly |
| OpenEvolve | An open-source evolutionary coding agent for discovering and optimizing algorithms |
| Kai | An autonomous agent that evolves codebases by finding and patching vulnerabilities |
| Sift | A private, local AI copilot for searching and organizing your files |
I co-founded the Institute for AI-Driven Discovery of Algorithms (AIDDA), I'm a dev ambassador at Qwen, and I'm part of the AIXI Labs community. I also used to lead the AI4Science community at alphaXiv.
