I build models, then ask where they fail.

I am a PhD student working at the intersection of artificial intelligence and life science, after earlier AI R&D roles at SenseTime and Baidu.

My background spans deep-learning model design, generative media, image restoration, multimodal systems and engineering deployment. That experience now informs work around protein representation, structure, antibody optimization and experiment-guided learning.

The through-line is the attempt to turn a powerful learned prior into a reliable decision system—one that remains aware of physics, data limits and real-world validation.