Asahi Ushio
Research Engineer at Google
London, England, United Kingdom
Summary
👤
Senior🎓
Top SchoolAsahi Ushio is a research engineer with nine years' experience building and shipping AI systems across industry and academia, currently working on state-of-the-art audio generation at Google (Lyria 2). He has a strong NLP and ML research background—PhD-level work at Cardiff and multiple internships and roles at Snap, Amazon, and Cogent Labs—bridging multilingual text, speech, and music models. He led development of kotoba-whisper, a Japanese ASR distilled from Whisper, and has published on temporal shifts in social media and counterfactual detection, showing both applied and theoretical strengths. Comfortable moving models into production, he has implemented pipelines for search-index evaluation and client-facing AI APIs. Fluent in English, Japanese and Chinese studies, he brings a rare combination of deep research, cross-lingual expertise, and production engineering to multimodal AI.
9 years of coding experience
5 years of employment as a software developer
Doctor of Philosophy - PhD, machine learning, natural language understanding, Doctor of Philosophy - PhD, machine learning, natural language understanding at Cardiff University / Prifysgol CaerdyddEnglish Language and Literature, General, English Language and Literature, General at University of WashingtonChinese Language and Literature, Chinese Language and Literature at Beijing Language and Culture UniversityMaster of Engineering - MEng, Machine Learning, Convex Optimization, Signal Processing, 4.8/5, Master of Engineering - MEng, Machine Learning, Convex Optimization, Signal Processing, 4.8/5 at Keio University
Japanese, English
Github Skills (64)
transformers10
cross-domain10
pytorch10
twitter10
named-entity-recognition10
python10
evaluation10
emoji10
machine-learning10
bart10
sentiment-analysis10
sentiment10
deep-learning10
computational-social-science10
question-answering10
Programming languages (2)
Github contributions (5)
asahi417/asahi417.github.io
Nov 2016 - Nov 2021
Contributions:63 pushes, 1 branch in 5 years
cardiffnlp/tweetnlp
May 2022 - Jan 2023
TweetNLP for all the NLP enthusiasts working on Twitter! The Python library tweetnlp provides a collection of useful tools to analyze/understand tweets such as sentiment analysis, emoji prediction, and named entity recognition, powered by state-of-the-art language models specialised on Twitter.
Contributions:1 release, 137 commits, 11 PRs in 8 months
pythonworking-onentitytwitter-sentiment-analysisstate-of-the-art
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.