Yihe is interested in large language models, geometric deep learning, and representation learning. She aims to bring a principled approach to deep learning, while maintaining scalability and robustness.

Prior to Princeton, Yihe has worked as a research engineer at Google and Microsoft.