Kyuyeun Kim
Software Engineer at Google
San Francisco Bay Area United States
Summary
🤩
Rockstar🎓
Top SchoolKyuyeun Kim is a software engineer specializing in ML accelerator performance and model optimization, currently focused on TPU performance at Google. With a PhD in Computer Science and prior experience building GPU drivers at LG and contributing quantization tooling at Qualcomm AI Research, he bridges low-level systems and practical deep learning inference improvements. He has hands-on experience optimizing LLM and transformer workflows and has contributed quantization implementations and tests to the widely used AIMET library. Based in the San Francisco Bay Area, he combines academic rigor in GPGPU systems with product-focused engineering at scale. Notably, his background spans both kernel/GPU driver development and high-level model compression, enabling unique end-to-end performance wins across hardware and software stacks.
4 years of coding experience
4 years of employment as a software developer
Doctor of Philosophy - PhD Computer Science, Doctor of Philosophy - PhD Computer Science at Ulsan National Institute of Science and Technology
Korean, English
Github Skills (11)
quantization10
machine-learning10
deep-learning10
tensorflow10
compress9
lossless-compression9
quants9
net9
pruning8
compression8
deep-neural-networks8
Programming languages (1)
Github contributions (2)
quic/aimet
Oct 2021 - Oct 2022
AIMET is a library that provides advanced quantization and compression techniques for trained neural network models.
Role in this project:
ML Engineer
ML EngineerContributions:1 review, 48 commits, 60 PRs in 11 months
Contributions summary:Kyuyeun's commits focus on the implementation of quantization techniques within the AIMET library. They removed a TensorFlow graph implementation and refactored code to use a connected graph for parameter quantization. Further work included simplifying parameter quantization logic, adding support for TF Transformer models, and implementing unit tests to validate these changes. These changes involve modifying core components for model optimization and inference acceleration.
pytorchtechniquesdeep-learningpruningcompression
quic-kyuykim/aimet
Mar 2022 - Oct 2022
AIMET is a library that provides advanced quantization and compression techniques for trained neural network models.
Contributions:20 pushes, 65 branches in 7 months
pytorchtechniquesdeep-learningcompressionmachine-learning
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.