Dan Fu builds algorithms that make machine learning models run faster on modern hardware. He is the VP of Kernels at Together AI and an incoming computer science professor at UCSD, best known for his foundational work on FlashAttention.
His research tackles the hidden bottlenecks in AI infrastructure. Whether breaking down the true cost of an AI request or optimizing GPU scheduling, he focuses on the systems-level mechanics that dictate inference performance. Dan completed his PhD at Stanford.