Noam Gat
Staff Software Engineer, Google Research
Ramat Gan, Tel-Aviv District, Israel
Summary
🤩
Rockstar🎓
Top SchoolNoam Gat is a Staff Software Engineer and technology leader with 17 years of experience building high-impact AI, real-time security, and consumer systems, currently focused on reinforcement learning research at Google Research. He has co-founded and scaled startups (Swapp, Tacticsoft), led R&D teams, and helped raise $20M while shipping AI-driven products that combine ML, computer vision, computational geometry and high-performance simulation. Noam’s work spans the full stack—from kernel-level optimizations handling multi-GB/s streams in the IDF to shader-level graphics and LLM inference improvements in notable open-source projects like Ogre3D and vllm. He’s skilled at transforming organizations into data-driven teams via instrumentation, pipelines and experimentation, and has tackled fairness in generative AI as well as large-scale reward modeling and RL-for-recommendation efforts. Known for hands-on engineering and strategic leadership, he blends deep systems expertise with product sensibility and a knack for turning complex 3D and ML challenges into practical, production-ready solutions.
18 years of coding experience
19 years of employment as a software developer
MSc, Computer Science, Machine Learning, MSc, Computer Science, Machine Learning at Tel Aviv UniversityBSc, Computer Science, Cognitive Science, BSc, Computer Science, Cognitive Science at The Hebrew University of Jerusalem
Stackoverflow
1reputation
0reached
0answers
0questions
Github Skills (28)
pytorch10
mesh10
c-language10
geometry10
gltf10
python10
rendering10
inference10
sampling10
glsl10
render10
llm10
geometry-shader10
transformer10
engine10
Programming languages (15)
Github contributions (5)
vllm-project/vllm
Oct 2023 - Aug 2025
A high-throughput and memory-efficient inference and serving engine for LLMs
Role in this project:
ML Engineer
ML EngineerContributions:2 reviews, 13 PRs, 45 comments in 1 year 10 months
Contributions summary:Noam primarily contributed to the VLLM project by implementing and improving core inference functionalities for LLMs. Their work involved adding a logits processor API to sampling parameters, enabling LM format enforcer guided decoding, and updating the flashinfer backend. These changes improve the flexibility and performance of the inference engine, particularly for JSON output and correct function calls using OpenAI's API interface. The user also fixed bugs and incorporated updates to support PagedAttention forwards, enhancing the project's compatibility and stability.
inferencegptllmpytorchmodel-serving
OGRECave/ogre
Aug 2008 - Feb 2010
high-performance rendering backend (C++, Python, C#, Java)
Role in this project:Full-stack Developer
Full-stack DeveloperContributions:53 commits in 1 year 6 months
Contributions summary:Noam primarily contributed to the Ogre3D engine, focusing on enhancing the rendering capabilities using geometry shaders and 'render to vertex buffer' techniques. The contributions include adding a particle system example (ParticleGS), incorporating geometry shader-based samples (Isosurf), and modifying shaders and rendering code related to existing samples. The user also updated the samples media directory and made improvements to GLSL geometry shader handling, demonstrating a deep understanding of the engine's rendering pipeline and shader programming. Further contributions show the porting and conversion of existing samples to the new sample framework and fixes for Linux compilation.
backendcppcsharpjavapython
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.