Dongwon Paek
ML Engineer at NAVER Corp
South Korea
Summary
👤
Senior🎓
Top SchoolDongwon Paek is an ML Engineer with eight years of hands-on experience building embedded and on-device AI systems, currently working on ML and MLOps at NAVER. He previously led embedded AI development at a Samsung C-Lab spin-out (루플) and contributed quantized kernel implementations and UINT8 support to Samsung/ONE’s on-device neural engine. His background includes freelance IoT and DSP projects—FFT-based signal processing, vision-based situational recognition, and wireless flight-controller systems—showing a strong mix of software, hardware, and systems-level ML. As an SW Maestro alumnus he prototyped an EMG-driven prosthetic using machine learning, reflecting applied research turned into products. He also shares technical writing in Korean through his IT blog and brings practical testing and deployment experience for production-constrained devices.
8 years of coding experience
3 years of employment as a software developer
인하대학교
English, Japanese, 인도네시아어
Github Skills (10)
neural-network10
compiler10
quantization10
machine-learning10
compiler-compiler10
optimization10
cprogramming-language9
c-language9
pytorch9
tensorflow8
Programming languages (5)
Github contributions (5)
Samsung/ONE
Jul 2020 - Nov 2020
On-device Neural Engine
Role in this project:
ML Engineer
ML EngineerContributions:31 reviews, 42 commits, 58 PRs in 3 months
Contributions summary:Dongwon primarily contributed to the `On-device Neural Engine` repository by implementing and improving machine-learning related functionalities. Their work involved adding and optimizing kernels, specifically for the Tanh and LogSoftmax activation functions, enabling support for UINT8 data types, and integrating the Maximum and Minimum kernels. They also added example implementations in PyTorch. The user's contributions focused on expanding the capabilities of the neural engine, including testing, to support quantized operations.
on-device-aion-deviceoptimizationruntimeneural-engine
underflow101/sos-practice
Dec 2020 - Apr 2021
Make open sources friendlier, Make people participate, Make people get involved into SOS.
Contributions:2 reviews, 20 commits, 17 PRs in 3 months
sossources
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.