Welcome to the UCSD Hao AI Lab website! We are passionate about designing strong, efficient, and secure machine learning models and algorithms, and in building scalable, practical distributed systems that can support real-world machine learning workloads. We also develop and maintain open-source models, evaluations, and systems to democratize the access of large models. We are affiliated with the UCSD ML System Group and UCSD NLP Group.

News

2025/04
NVIDIA gifted us a DGX B200! See our recent post here.
2025/03
Our research paper DistServe is adopted by NVIDIA Dynamo and highlighted in GTC'25!
2025/01
We released the FastVideo project with FastHunyuan and FastMochi models to accelerate video DiTs.
2024/12
🎉🎉 In 2024, We have 1 ICLR, 1 OSDI, 6 ICML, 1 ACL, 1 COLM papers, 1 EMNLP, and 2 NeurIPS papers. Kudos to our lab members for a productive year!
2024/07
Our lab members Lanxiang Hu, Yichao Fu, Runyu Lu will be attending ICML'24 to present their work: OSD, MuxServe, Lookahead Decoding, and CLLMs.
2024/05
Six papers are accepted to ICML 24! Check out: OSD, MuxServe, Lookahead Decoding, Chatbot Arena, APISERVE, CLLMs.
2024/03
DistServe accepted to OSDI 24! 🎉
2023/07
Hao officially joined UCSD!

 

Sponsors

Hao AI Lab is sponsored by HDSI, NVIDIA, Google, and Snowflake. We are grateful for their support!