Jeongseok Kang
Software Engineer at Lablup Inc.
Seoul, South Korea
Summary
👤
Senior🎓
Top SchoolJeongseok Kang is a software engineer based in Seoul with a decade of hands-on experience building backend systems, test automation, and ML-related CUDA integrations. Currently at Lablup, he contributes to Backend.AI where he improved integration tests and enhanced compute-session monitoring for a container-based ML platform supporting diverse accelerators. His open-source work includes CUDA-to-PyTorch refactors in the widely used bitsandbytes library, improving compatibility for k-bit quantized LLM workflows. He combines academic rigor from a computer engineering master's at Korea Aerospace University with practical product experience from Android and backend internships. Colleagues rely on him to bridge low-level GPU/compute details with resilient backend testing and tooling. He’s equally comfortable diving into CUDA/PyTorch nuances as he is strengthening CI and integration test suites.
10 years of coding experience
1 year of employment as a software developer
Bachelor's degree, Software Engineering, Bachelor's degree, Software Engineering at Korea Aerospace University석사, 컴퓨터공학, 4.3, 석사, 컴퓨터공학, 4.3 at 한국항공대학교
Github Skills (17)
json10
pytorch10
python10
testing10
machine-learning10
llm10
cuda10
quantization10
test-automation10
command-line-interface9
command-line9
cli9
distributed-computing8
dockers7
api7
Programming languages (15)
Github contributions (5)
lablup/backend.ai
Jun 2022 - Jan 2023
Backend.AI is a streamlined, container-based computing cluster platform that hosts popular computing/ML frameworks and diverse programming languages, with pluggable heterogeneous accelerator support including CUDA GPU, ROCm GPU, TPU, IPU and other NPUs.
Role in this project:
Back-end Developer & Test Automation Engineer
Back-end Developer & Test Automation EngineerContributions:131 reviews, 102 commits, 130 PRs in 7 months
Contributions summary:Jeongseok focused on updating and improving integration tests for the Backend.AI platform. They modified test assertions to use the JSON output of mutation commands, ensuring tests aligned with the project's structure. Furthermore, they updated existing tests related to keypair, scaling group, and user functionalities. The user also contributed to improving the test environment and adding commands to monitor compute sessions, enhancing the platform's testing coverage and reliability.
diversepythonpluggablebackendaicloud-computing
bitsandbytes-foundation/bitsandbytes
Apr 2023 - Aug 2023
Accessible large language models via k-bit quantization for PyTorch.
Role in this project:ML Engineer
ML EngineerContributions:3 reviews, 3 PRs, 14 comments in 4 months
Contributions summary:Jeongseok's contributions primarily focused on improving the CUDA setup and integration within the `bitsandbytes` library. They addressed typos, replaced CUDA runtime library calls with PyTorch APIs, and updated code to get CUDA version and device capabilities through PyTorch. These changes indicate a focus on ensuring the library's compatibility and efficient utilization of CUDA for accelerating machine learning tasks, specifically large language model quantization.
cudapytorch8-bitgpupruning
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.