Kenshin Abe
Software Engineer at Preferred Networks, Inc.
Chiyoda, Japan
Summary
🤩
Rockstar🎓
Top SchoolKenshin Abe is a software engineer with nine years' experience blending machine learning research and product-focused engineering, currently at Preferred Networks in Tokyo. A Kaggle Grandmaster, he brings deep expertise in graph neural networks and hyperparameter optimization—contributing backend features and robustness improvements to Optuna and implementing sparse RelGCN/GIN support in Chainer Chemistry. His background spans web API work at LINE and data analysis at Alibaba.com, giving him practical full-stack and data-savvy instincts. With a Master's in Computer Science from the University of Tokyo, he moves fluidly between research-grade model implementations and production engineering, often surfacing subtle tooling and testing refinements that improve reliability.
9 years of coding experience
2 years of employment as a software developer
University of Tokyo
Github Skills (11)
hyperparameter-optimization10
machine-learning10
deep-learning10
graph-convolutional-networks10
python10
optuna10
chainer10
testing9
sparse-matrix8
biology4
chemistry4
Programming languages (5)
Github contributions (5)
optuna/optuna
Dec 2021 - Jan 2023
A hyperparameter optimization framework
Role in this project:
Back-end Developer
Back-end DeveloperContributions:171 reviews, 182 commits, 73 PRs in 1 year
Contributions summary:Kenshin's contributions primarily revolved around enhancing the Optuna hyperparameter optimization framework. They focused on adding features to the `Study` class, such as the ability to enqueue trials with pre-defined parameter values and specifying custom attributes for trials. Furthermore, the user addressed code quality by fixing typos in the documentation and refactoring existing documentation. Their work also included adding functionality to the `RetryFailedTrialCallback` to manage failed trials and associated tests.
pythonoptimization-frameworkparallelhyperparameteroptimization
chainer/chainer-chemistry
Sep 2019 - Sep 2019
Chainer Chemistry: A Library for Deep Learning in Biology and Chemistry
Role in this project:ML Engineer
ML EngineerContributions:52 commits, 1 PR, 24 comments in 16 days
Contributions summary:Kenshin primarily focused on adapting and supporting the RelGCN and GIN models within the Chainer Chemistry library. They implemented sparse versions of these models and adapted dataset classes to support them. Further contributions included refactoring and adding support for GIN models, indicating a focus on graph neural network implementation and integration within the library.
chemistrygraph-convolutional-networkspythondeep-learningbiology
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.