Güray Özen
Machine Learning Compiler Engineer at NVIDIA
Switzerland
Summary
🤩
Rockstar🎓
Top SchoolGüray Özen is a Machine Learning Compiler Engineer with 11 years of experience building and optimizing compilers for GPU-accelerated workloads, currently at NVIDIA after a recent stint at Google. He holds a PhD in Computer Architecture and has a strong track record in MLIR-based compilers, contributing to high-profile open-source projects like IREE and LLVM (notably the NVGPU dialect and CUDA codegen). His work spans GPU code generation, tensor memory access optimizations, auto-parallelization, and device mapping interfaces that unlock real-world performance on CUDA devices and HPC systems. Past roles include research and engineering positions at Barcelona Supercomputing Center, IBM, and multiple teams at NVIDIA, where he influenced OpenMP/OpenACC language evolution. Güray combines deep research instincts with production-grade engineering, often bridging tutorial and API work (Python/C API) to make advanced GPU features more accessible. Colleagues describe him as a pragmatic optimizer who turns compiler theory into tangible speedups on top-tier hardware.
11 years of coding experience
9 years of employment as a software developer
UPC Universitat Politècnica de CatalunyaBachelor of science Computer Engineering, Bachelor of science Computer Engineering at Dokuz Eylul University
Stackoverflow
405reputation
21kreached
6answers
15questions
Github Skills (16)
cuda10
compiler10
c-language10
compiler-compiler10
gpgpu10
cprogramming-language10
gpu10
python10
mlr10
llvm10
machine-learning9
api8
apidoc8
wcf6
jit6
Programming languages (5)
Github contributions (5)
llvm/llvm-project
Jan 2023 - Jan 2023
The LLVM Project is a collection of modular and reusable compiler and toolchain technologies.
Role in this project:
Back-end Developer
Back-end DeveloperContributions:273 reviews, 2 commits, 214 PRs in 1 day
Contributions summary:Güray primarily contributed to the development of the NVGPU dialect within the LLVM project, specifically focusing on Python bindings and related C API implementations. Their work involved adding support for the `TensorMapDescriptorType` in the Python bindings and simplifying TMA (Tensor Memory Access) IR generation. They also implemented tutorial codes focusing on the nvgpu dialect and its advanced features. The user's contributions facilitated enhanced programmability and utilization of NVGPU hardware features through the LLVM compiler infrastructure.
compilerstechnologiesclangsubmittoolchain
iree-org/iree
Sep 2022 - Jan 2023
A retargetable MLIR-based machine learning compiler and runtime toolkit.
Role in this project:Backend Developer
Backend DeveloperContributions:44 reviews, 24 commits, 43 PRs in 3 months
Contributions summary:Güray primarily contributed to the IREE compiler, focusing on GPU code generation. They implemented features and addressed issues related to the transformation and optimization of code for GPU execution. The commits involved integrating new GPU transform dialects and device mapping interfaces, improving the generation of CUDA code. The user also worked on optimizing the code for CUDA devices, specifically relating to the settings of thread block sizes and vectorization.
mlirspirvvulkantensorflowcompiler
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.