Senior MLOps Engineer at Turing Inc. — building the data & ML infrastructure behind end-to-end autonomous driving.
I work at the intersection of MLOps, large-scale data pipelines, and vision-language models. Currently I architect distributed data pipelines and ML infrastructure for autonomous driving model development, spanning Databricks Delta Live Tables, on-prem NVIDIA H100 / Slurm clusters, and AWS Batch.
- ML infrastructure for autonomous driving — petabyte-scale driving-log pipelines feeding end-to-end / VLA model training
- Semantic video search over terabytes of driving logs, using NVIDIA Cosmos-Embed1 and Databricks Vector Search
- MOMIJI 🍁 — led the curation and publication of a web-scale Japanese image-text dataset for vision-language models
| Turing Inc. (2024–) | Senior MLOps Engineer — AD/VLA data pipelines, ML infra, VLM development |
| CyberAgent (2022–2024) | ML Engineer — CVR prediction for video ads on ABEMA |
| Cookpad (2016–2022, Tokyo & Bristol 🇬🇧) | Server-side / Research Engineer — search quality, ML for recipe data, NLP research |
Education: M.S. in Information Science and Technology (Machine Learning), The University of Tokyo · B.E. in Computer Science, Waseda University
- Yamaguchi, Inuzuka, Hiramatsu, Harashima. Non-ingredient Detection in User-generated Recipes using the Sequence Tagging Approach — W-NUT @ EMNLP 2020
- Ito, Inuzuka, Yamada, Harashima. Real World Voice Assistant System for Cooking — INLG 2019
- Inuzuka, Ito, Harashima. Step or Not: Discriminator for The Real Instructions in User-generated Recipes — W-NUT @ EMNLP 2018
🚴 Off the keyboard
Road cyclist — you'll find me climbing hills on weekends.