Hanbin Yoon
Software Engineer at Google
Mountain View, California, United States
Summary
🤩
Rockstar🎓
Top SchoolHanbin Yoon is a software engineer based in Mountain View with five years of professional experience and a deep background in systems and ML runtime engineering. Currently at Google, he contributes to high-performance back-end infrastructure, bringing prior R&D experience from HP Labs and Samsung Electronics. His open-source work on TensorFlow Runtime and the XLA compiler highlights practical contributions to core kernels and GPU GEMM execution paths, including implementing corert.while and BEF-based GEMM thunks. Hanbin combines low-level performance tuning with production-minded design, adding integer and fallback tensor support to improve runtime robustness. Trained at Cambridge and Carnegie Mellon in information and computer engineering, he blends academic rigor with hands-on systems craftsmanship. Colleagues would notice his knack for quietly refactoring complex compiler/runtime code to unlock measurable performance and correctness gains.
5 years of coding experience
5 years of employment as a software developer
Bachelor of Arts (B.A.), Information and Computer Engineering, Bachelor of Arts (B.A.), Information and Computer Engineering at University of CambridgeMaster of Science (M.S.), Electrical and Computer Engineering, Master of Science (M.S.), Electrical and Computer Engineering at Carnegie Mellon University
English
Github Skills (14)
kernel10
xla10
compiler10
gpu-programming10
c-language10
tensorflow10
cprogramming-language10
mlr10
runtime-environment10
control-flow10
performance-optimization9
cudnn8
ccl7
machine-learning6
Programming languages (2)
Github contributions (5)
tensorflow/runtime
Sep 2020 - Aug 2022
A performant and modular runtime for TensorFlow
Role in this project:
Back-end Developer
Back-end DeveloperContributions:2 reviews, 74 commits in 1 year 11 months
Contributions summary:Hanbin primarily contributed to the core runtime of TensorFlow by implementing kernels for the `corert` runtime environment. They developed the `corert.while` kernel, enabling loop functionality within the runtime. Their work also involved adding support for integer data types and fallback tensors within the `corert.while` kernel, showcasing a focus on improving the robustness and functionality of the TensorFlow runtime.
runtimeperformantmodulartensorflow
openxla/xla
Apr 2021 - Aug 2022
A machine learning compiler for GPUs, CPUs, and ML accelerators
Role in this project:Back-end & MLOps Engineer
Back-end & MLOps EngineerContributions:147 commits in 1 year 4 months
Contributions summary:Hanbin's contributions primarily focused on enhancing the XLA compiler, specifically related to GPU support. They implemented unit tests to exercise gemm thunks and refactored the code to improve GemmThunk instantiation. Furthermore, the user introduced and refined BefThunk, a feature utilizing BEF execution to perform GEMM operations via MLIR lowering. Additionally, they contributed to code refactoring and introduced various improvements related to the implementation of various XLIR kernels.
compilercommunity-drivenmachine-learningmodular
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.