Ryokan Ri
Research Engineer at Google DeepMind
Bunkyō, Japan
Summary
👤
Senior🎓
Top SchoolRyokan Ri is a research engineer based in Bunkyō, Japan, with a decade of experience focused on natural language processing and large language models. He combines deep academic training (PhD in Information Technology from the University of Tokyo) with production-oriented research roles across Studio Ousia, LINE, SB Intuitions, and now Google DeepMind. His work on knowledge-enhanced models and multilingual pretraining—illustrated by contributions to the LUKE repository, adding sentence tokenizers and a multilingual entity vocabulary—bridges linguistic insight and scalable engineering. Ryokan has published in top NLP venues and brings practical experience from industry internships at IBM and Studio Ousia, plus a background in language teaching that informs his approach to language data. He’s particularly adept at making research codebase-ready through clean type hinting and multilingual adaptations, enabling models to move from prototype to robust, multilingual deployments.
10 years of coding experience
3 years of employment as a software developer
Doctor of Philosophy - PhD, Information Technology, Doctor of Philosophy - PhD, Information Technology at 東京大学The University of Sydney
English, Chinese, Japanese
Github Skills (7)
nlp10
multilingual10
python10
natural-language-processing10
machine-learning9
data-structure7
data-structures7
Programming languages (6)
Github contributions (5)
studio-ousia/luke
Apr 2020 - Oct 2022
LUKE -- Language Understanding with Knowledge-based Embeddings
Role in this project:
Back-end Developer
Back-end DeveloperContributions:285 commits, 18 PRs, 120 pushes in 2 years 6 months
Contributions summary:Ryokan primarily contributed to the implementation of the "LUKE" project, focusing on the addition of sentence tokenizers and a multilingual entity vocabulary. These changes involved modifying the code to support multiple languages during the pretraining phase. The contributions also include adapting the code for multilingual training and adding type hinting for improved code maintainability.
language-understanding
xuzf-git/entity-diambiguation-luke
Dec 2020 - Jul 2021
LUKE -- Language Understanding with Knowledge-based Embeddings
Contributions:341 commits in 7 months
language-understanding
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.