Kichang Kim is a Tokyo-based programmer with a decade of engineering experience, combining a strong control-systems academic background from Tokyo Institute of Technology with hands-on software development at NHN PlayArt. He contributes to open-source ML tooling—notably enhancing DeepDanbooru by enabling GPU-accelerated Grad-CAM and adding experimental ResNet variants—showing a focus on model performance and architecture experimentation. His blend of mechanical/control engineering training and practical ML/engineering work gives him a systems-oriented approach to building performant software. Having started his professional life in public service as a librarian, he brings patience and attention to detail to collaborative projects.
10 years of coding experience
Master's degree, Department of Mechanical and Control Engineering, Master's degree, Department of Mechanical and Control Engineering at 東京工業大学 / Tokyo Institute of Technology
AI based multi-label girl image classification system, implemented by using TensorFlow.
Role in this project:
ML Engineer
Contributions:4 releases, 56 commits, 8 PRs in 3 years 1 month
Contributions summary:Kichang primarily focused on enhancing the deepdanbooru image classification system, particularly by enabling GPU usage for the Grad-CAM feature, improving its performance. They also added a new ResNet custom model version (v3) for experimental use, indicating a focus on model architecture and experimentation. Furthermore, the user made changes to package naming conventions and core functionality.
Contributions:1 release, 43 commits, 1 PR in 5 years 3 months
downloaderdanbooruimage-downloader
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.