Huilin Tai

Hello! I am a first-year PhD student in Computer Science at Rice University, advised by Prof. Andrew H. Song in the APEX Lab. I am also affiliated with MD Anderson Cancer Center.

I completed a thesis-based M.S. in Computer Science at Columbia University, working with Prof. Mohammed AlQuraishi on evaluating representations from long-context genome language models. I received my B.Sc. in Statistics and Computer Science from McGill University, where I worked with Prof. Jun Ding, Prof. Anmar Khadra, and Prof. Hamed Hatami.

My research focuses on interpreting, evaluating, and controlling multimodal foundation models, with cancer research as a major application. I pursue two connected directions:

Interpretability, evaluation, and control. I study how models represent and select evidence, how their behavior changes under interventions, and what these observations can actually tell us about the mechanisms behind their predictions.

Multimodal foundation models for cancer. Histology, molecular measurements, and clinical data offer different, incomplete views of tumor biology. I am interested in how models can bring these views together to understand tissue organization, cellular interactions, and treatment response—and in distinguishing what can be inferred from the available evidence from what remains unresolved.

I am always happy to discuss research ideas or collaborations. Feel free to reach out! Outside of research, I enjoy squash, sketching, and writing.