prof_pic.jpg

Jiawei Zhao

I am a Research Scientist at Meta AI (FAIR & GenAI), where I work on topics related to LLM reasoning and efficiency. I received my Ph.D. at Caltech, where I invented GaLore, a hardware-efficient optimizer for LLMs.

My research interests include 1) scalable and efficient optimization algorithms for LLM training, 2) efficient approaches for LLM reasoning and large-scale reinforcement learning, 3) efficient long-context LLM inference. Check out my research for more details.

Reach me at jwzhao at meta dot com (work) or jiawei at caltech dot edu (personal).

news

Jul 21, 2025 We are organizing the first workshop on Efficient Reasoning at NeurIPS 2025 and are calling for papers. See you in San Diego!
Jun 01, 2025 I gave a talk at TL;DR’25 at Rice University about my recent work on Hardware-Efficient Learning Algorithms for Large Language Models.
May 13, 2024 I gave a talk at MLSys’24 about my recent work on Memory-Efficient LLM Training.
May 02, 2024 GaLore is accepted by ICML 2024 as Oral!