Seonghyeon Drew
조교수 at 경북대학교(Kyungpook National University)
Daegu, South Korea
Summary
👤
Senior🎓
Top SchoolSeonghyeon Drew is an assistant professor and machine learning researcher based in Daegu with 11 years of experience bridging academia and industry. After earning a PhD in Computer Engineering from POSTECH, he progressed from research roles and industry positions (ScatterLab, NAVER, Kakao) to a postdoc and now a faculty appointment at Kyungpook National University. His work focuses on graph neural networks and practical PyTorch implementations—he contributed notable improvements to a Graph Attention Network repo, adding sparse support, better initialization, and dropout and backprop fixes. Comfortable moving models from research to production, he combines deep theoretical training with hands-on engineering across distributed systems and web stacks. Colleagues know him for clear, reproducible implementations and an ability to spot subtle performance issues that improve real-world model reliability.
11 years of coding experience
3 years of employment as a software developer
Ph. D., Computer Engineering, 3.66/4.3, Ph. D., Computer Engineering, 3.66/4.3 at 포항공과대학교
Korean, English
Github Skills (9)
neural-network10
attention-mechanism10
pytorch10
machine-learning10
python10
dropout9
sparse-data9
sparse-matrix9
backpropagation8
Programming languages (7)
Github contributions (5)
Diego999/pyGAT
Sep 2018 - Oct 2018
Pytorch implementation of the Graph Attention Network model by Veličković et. al (2017, https://arxiv.org/abs/1710.10903)
Role in this project:
ML Engineer
ML EngineerContributions:6 commits, 1 PR, 16 comments in 1 month
Contributions summary:Seonghyeon primarily focused on implementing and refining the Graph Attention Network (GAT) model using PyTorch. Their contributions included supporting sparse versions of the GAT, changing weight initialization methods for improved performance, and adding a dropout feature to prevent overfitting. Additionally, they differentiated between dense and sparse versions of the model and fixed backpropagation related issues.
pytorchthe-graphpythonarxivabs
sh0416/oommix
Nov 2020 - May 2021
Contributions:79 commits, 1 push in 6 months
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.