Summary
Junru Shao is a Distinguished Engineer based in San Francisco with nine years of MLSys experience focused on ML frameworks, distributed training/inference, and ML compilers. A top-1 contributor to TVM since 2021 and serving on its Apache PMC, he blends deep open-source leadership with industrial production experience across OctoML, AWS, OpenAI, and now NVIDIA. He has led technical teams and managed delivery of large-scale pretraining and compiler tooling, repeatedly moving research-grade optimizations into production. Junru’s background—MS in Machine Learning from Carnegie Mellon and early contributions to MXNet—reflects a strong mix of academic rigor and systems engineering; less obvious is his knack for balancing committer-level open-source work with hands-on ML system tuning at scale.
9 years of coding experience