Masashi Shibata
Engineering Manager at Preferred Networks, Inc.
Chiyoda, Japan
Summary
🤩
Rockstar🎓
Top SchoolMasashi Shibata is an engineering manager based in Chiyoda, Japan with nine years of experience building high-performance ML systems, production services, and developer tools. He blends hands-on systems work (Python, C/C++, Cython, gRPC, Kubernetes) with leadership at Preferred Networks after a technical career at CyberAgent where he optimized inference servers handling hundreds of thousands RPS and won internal engineering awards. An active open-source contributor and author, he created go-prompt and kube-prompt, is an Optuna committer and Kubeflow/Katib reviewer, and has implemented GPU-accelerated special functions for prominent libraries like CuPy and Chainer. He’s comfortable shipping low-latency distributed services and tight C/C++–Python integrations, and often bridges research and production—evident from NeurIPS/Optuna competition work and CUDA kernel contributions.
9 years of coding experience
8 years of employment as a software developer
Information Technology, Information Technology at 明石工業高等専門学校
Github Skills (18)
python10
chainer10
machine-learning10
distributions10
kernel10
deep-learning10
f10
kernel-mode10
cupy10
cuda10
probability-distribution10
scientific-computing10
c-language9
numpy9
beta-distribution9
Programming languages (6)
Github contributions (5)
cupy/cupy
May 2018 - Nov 2018
NumPy & SciPy for GPU
Role in this project:
Back-end Developer
Back-end DeveloperContributions:216 commits, 43 PRs, 10 pushes in 5 months
Contributions summary:Masashi contributed several special mathematical functions to the CuPy library, implementing their CUDA kernels. They added functions such as `gammaln`, `digamma`, `gamma`, `zeta`, and `polygamma`. These changes involved defining the kernel code, including relevant definitions, and integrating them within the library. The user also fixed PEP8-related style issues, ensuring code quality and consistency.
cudapythoncusolvergpunumpy
chainer/chainer
May 2018 - Feb 2019
A flexible framework of neural networks for deep learning
Role in this project:ML Engineer
ML EngineerContributions:200 commits, 55 PRs, 17 pushes in 9 months
Contributions summary:Masashi's contributions primarily focused on the implementation of mathematical functions and the development of distribution classes within the Chainer framework for deep learning. They added functionality for log gamma, digamma, and polygamma functions, including CUDA implementations for GPU acceleration. Additionally, the user introduced and tested a variety of probability distributions such as Bernoulli, Beta, and Normal, and added code for a multivariate normal distribution.
cudapythonmxnetcaffe2flexible-framework
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.