Shixiang Gu
Senior Staff Research Scientist at Google DeepMind
San Francisco, California, United States
Summary
🤩
Rockstar🎓
Top SchoolShixiang Gu is a senior staff research scientist with 10+ years building cutting-edge ML systems across deep learning, reinforcement learning, robotics, and large language model post-training. Currently leading multilingual efforts at Google DeepMind (and founder of its Japan LM team), he has driven Gemini models to top-ranked non-English performance and spearheaded "Gemini for Japan" industry collaborations. His work spans research and applied production—co-authoring major reports (including GPT-4 and Gemini) and contributing practical tooling such as multilingual and evaluation enhancements to high-profile open-source eval frameworks. A Cambridge PhD co-supervised by Turner, Ghahramani and Schölkopf, he combines rigorous probabilistic ML foundations with hands-on robotics and production ML experience from Google Brain, OpenAI, and industry internships. Multilingual and globally mobile, he leverages lived experience across six countries and three written languages to inform cross-cultural product launches and research impact.
10 years of coding experience
13 years of employment as a software developer
Doctor of Philosophy (Ph.D.) Machine Learning, Doctor of Philosophy (Ph.D.) Machine Learning at University of CambridgeBachelor of Applied Science (B.A.Sc.) Engineering Science - Electrical and Computers Option Minor in Robotics & Mechatronics, Bachelor of Applied Science (B.A.Sc.) Engineering Science - Electrical and Computers Option Minor in Robotics & Mechatronics at University of Toronto
Japanese, Chinese, English
Github Skills (4)
data-generation10
yaml10
python10
openai-api8
Programming languages (3)
Github contributions (5)
openai/evals
Mar 2023 - May 2023
Evals is a framework for evaluating LLMs and LLM systems, and an open-source registry of benchmarks.
Role in this project:
Full-stack Developer
Full-stack DeveloperContributions:6 reviews, 37 PRs, 33 pushes in 2 months
Contributions summary:Shixiang contributed to the `openai/evals` repository by adding multilingual support and examples, generating test data and YAML configuration files for model-graded evaluations. They modified scripts related to data generation and evaluation, including scripts for battle and model-graded generation. Additionally, they introduced changes to the logging and recording functionalities of the evaluation framework, including the introduction of pause/unpause recording functionality. These changes suggest contributions spanning across the evaluation framework and potentially involving language model processing and model comparison.
rlbayes/rllabplusplus
Mar 2017 - Jul 2017
Contributions:2 commits, 2 pushes, 1 branch in 3 months
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.