Dan Fu: FlashAttention Co-Author at Together AI | Generated by AI

Home 2026.06

Question: Who is Dan Fu, the VP of Kernels at Together AI?

Answer:

Dan Fu is essentially Tri Dao’s close collaborator and the person running Together AI’s low-level GPU engineering. Here’s the breakdown:

Who he is:

Assistant Professor at UCSD (CSE department), leading the SandyResearch Lab, and VP of Kernels at Together AI — building a world-class team focused on low-level performance engineering and GPU kernels.

His research develops efficient algorithms for machine learning, with a focus on solutions that are both theoretically efficient and practically fast on modern hardware.

His key work:

The model he operates under:

Dan Fu runs a lab at UCSD focused on higher-risk fundamental research, including his personal passion project on FFT algorithms. Together AI co-founder Tri Dao is at Princeton. The model is symbiotic: de-risk ideas in academia, productionize them at Together AI. PhD students join the company, Together AI interns work on longer-term research in academic labs.

So the Together AI kernels org is essentially a Stanford/Princeton systems mafia productionizing their own research — FlashAttention, Mamba, H3, ThunderKittens — all flowing from academia directly into the inference stack. Very similar to how DeepSeek operates: research-first, implementation depth as the moat.

References:


Back Donate