Masahiro Sakai
Data Structure & Algorithm Engineer
Kawasaki, Japan
Summary
👤
Senior🎓
Top SchoolMasahiro Sakai is a Data Structure & Algorithm Engineer with 24 years of experience blending theoretical computer science research and hands-on software development. He has driven engineering and management at Preferred Networks and now researches at Noeon Research, applying formal methods, SAT/SMT solving, and optimization to real-world systems. His background includes SMT-based test generation, bounded model checking, and contributions to high-profile open-source ML libraries—most notably extending CuPy’s SciPy GPU capabilities and improving ONNX export in Chainer. A Keio graduate who co-translated seminal PL texts into Japanese, he pairs deep expertise in programming language theory and category theory with pragmatic engineering discipline. Colleagues describe him as relentlessly curious and perfection-minded, often turning abstract math into production-quality code.
24 years of coding experience
19 years of employment as a software developer
Master of Media and Governance, Cyber Informatics, Master of Media and Governance, Cyber Informatics at 慶応義塾大学 / Keio University
Japanese, English, German, Russian
Stackoverflow
93reputation
2kreached
0answers
2questions
Github Skills (18)
python10
chainer10
machine-learning10
onnx10
numpy10
deep-learning10
cupy10
cython10
scipy9
cusolver9
gpu9
neural-network9
cuda9
testing8
linear-algebra8
Programming languages (21)
Github contributions (5)
chainer/chainer
Oct 2019 - Mar 2020
A flexible framework of neural networks for deep learning
Role in this project:
ML Engineer
ML EngineerContributions:29 commits, 16 PRs, 14 comments in 4 months
Contributions summary:Masahiro primarily contributed to the ONNX-Chainer integration for the Chainer deep learning framework. Their work included fixing exporters for specific functions like `Separate` and adding a new exporter for `Permutate`, enabling broader ONNX model conversion capabilities. They also addressed documentation issues and implemented fixes to ensure correct behavior of the ONNX conversion process and handling edge cases within the generated ONNX models. Furthermore, the user added a converter for n_step_gru function.
deep-learningneural-networkpythonneural-networksmachine-learning
cupy/cupy
Feb 2019 - Mar 2019
NumPy & SciPy for GPU
Role in this project:Back-end Developer & Test Automation Engineer
Back-end Developer & Test Automation EngineerContributions:23 commits, 4 PRs, 12 comments in 1 month
Contributions summary:Masahiro primarily contributed to the CuPy library by fixing bugs, implementing new functionalities, and enhancing the existing codebase. They addressed memory leaks related to Cython classes and corrected data type issues within linear algebra functions. Additionally, the user implemented LU decomposition and solve functionalities within the CuPy's SciPy integration, demonstrating an ability to extend the library's capabilities. Furthermore, the user was involved in the test suite updates.
gpunumpyscipycudacudnn
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.