Yineng Zhang
Senior Director at Together AI
San Francisco, California, United States
Summary
🤩
Rockstar🎓
Top SchoolYineng Zhang is a Senior Director leading the inference team at Together AI, bringing six years of focused experience in production ML systems and model-serving optimization. He has progressed from hands-on core developer work on SGLang—where he implemented InternLM2 support, integrated FlashInfer, and improved test and build stability—to leading inference strategy and execution at multiple startups. Yineng combines deep backend engineering (GPU and compilation-aware optimizations) with cross-cutting research contributions showcased in MLSys and ACM venues, and has spoken at industry events including AMD AI Dev Day and PyTorch Conference. Based in San Francisco, he has driven model performance teams at Baseten and large-scale ML infrastructure at Meituan and Baidu, consistently turning research into deployable inference solutions. Colleagues describe him as pragmatic and curious—his GitHub bio “Just for fun 🌁” hints at a playful approach to complex systems engineering.
7 years of coding experience
7 years of employment as a software developer
Bachelor of Engineering, ENGINEERING, Bachelor of Engineering, ENGINEERING at Jiangnan University
Chinese, English
Github Skills (11)
code-optimization10
pytorch10
integrate10
python10
integrations10
data-integration10
deep-q-learning9
deep-learning9
cuda9
c-language5
cprogramming-language5
Programming languages (13)
Github contributions (5)
sgl-project/sglang
Jul 2024 - Mar 2026
SGLang is a high-performance serving framework for large language models and multimodal models.
Role in this project:
Back-end Developer
Back-end DeveloperContributions:10 releases, 1437 reviews, 2935 PRs in 1 year 8 months
Contributions summary:Yineng implemented support for the InternLM2 model within the SGLang framework, indicating a focus on expanding the framework's capabilities to support various language models. They added a pre-commit configuration to improve code quality. The user also added unit tests and corrected the code for the MTP in a follow-up commit, improving stability. They also integrated with FlashInfer and addressed potential issues during compilation, showcasing a focus on optimization and compatibility.
large-language-modelsmultimodalcudainferencellama
zhyncs/zhyncs.github.io
Aug 2019 - Jun 2026
Contributions:50 pushes, 1 branch in 6 years 11 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.