Mingheng Wu 吴明恒

I am a Research Scientist at ByteDance Seed, working on systems for large-scale language models. I build and optimize systems across the LLM stack: pretraining (PyTorch-native graph compilation, ND parallelism), inference (heterogeneous serving), and system performance modeling. I am a contributor to PyTorch and torchtitan.

I received my dual M.S. from the University of Washington and Tsinghua University, where I was advised by Yunhao Liu. Previously, I received my B.Eng. in Computer Science from Tsinghua University.