Hyukjin Jeong
Staff Software Engineer at Samsung Electronics
Seoul, United States
Summary
👤
Senior🎓
Top SchoolHyukjin Jeong is a Staff Software Engineer based in Seoul with eight years of experience building on-device NPU toolchains and compiler-backed neural network quantization. At Samsung Electronics he led SoC architecture and NPU toolchain work, shipping quantization techniques (integer-only, LLM quantization, Microscaling) and compiler optimizations for commercial NPU products (NQ AI Gen2/Gen3). He has full-stack expertise across quantization, frontends (Executorch, ONNX, TFLite) and compiler passes such as layer fusion, pinning, pipelining, and memory allocation that enabled 20+ optimized AI applications including a transformer-based live translator. His open-source contributions to Samsung/ONE focus on expanding tflchef operation support (Rsqrt, Tanh, Tile) and adding unit tests, reflecting attention to correctness as well as capability growth. With a PhD from Seoul National University and research roots in distributed compiler systems, he pairs deep academic rigor with pragmatic production engineering. An underappreciated strength is his track record of turning research ideas into deployable on-device ML features, from LLM PoCs to TV-integrated NPU products.
8 years of coding experience
Doctor of Philosophy - PhD, Electrical and Computer Engineering, Doctor of Philosophy - PhD, Electrical and Computer Engineering at 서울대학교 (Seoul National University)Bachelor's degree, Electircal and Computer Enginnering, Bachelor's degree, Electircal and Computer Enginnering at Seoul National University
Korean, English
Github Skills (6)
neural-network10
tflite10
c-language10
cprogramming-language10
python10
tensorflow9
Programming languages (4)
Github contributions (5)
Samsung/ONE
Apr 2020 - Jan 2023
On-device Neural Engine
Role in this project:
Back-end Developer
Back-end DeveloperContributions:2951 reviews, 629 commits, 1483 PRs in 2 years 9 months
Contributions summary:Hyukjin's commits primarily focus on enhancing the tflchef module within the "On-device Neural Engine" repository. Their contributions include implementing support for the Rsqrt and Tanh operations, as well as introducing a unit test for the constructor of the Rsqrt operation. The user has also added support for the Tile operation, showcasing their involvement in expanding the capabilities of the neural engine to handle a broader range of operations.
on-device-aion-deviceoptimizationruntimeneural-engine
jinevening/ONE
May 2020 - May 2023
On-device Neural Engine
Contributions:561 pushes, 1464 branches in 2 years 11 months
on-deviceneural-enginemachine-learningneural-networktensorflow
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.