Design, build, maintain and optimize the ML Platform systems and tools for perception, prediction, and planner development, enabling ML engineers to iterate on dataset curation, ML modeling, training, evaluation and deployment into the AD/ADAS stack shipped in Toyota vehicles
Develop user-friendly tooling, frameworks and libraries to support end-to-end ML engineering (model tracking, performance metrics, failure-mode introspection)
Build and maintain efficient dataset generation, cloud training and evaluation pipelines; optimize data loading and distributed training, plus cloud/edge deployment
Review code and collaborate with ML/ML Platform engineers to accelerate incremental improvements; contribute to long-term platform strategy
Operate in a high-velocity, agile environment with a hybrid Tokyo office presence ( Nihonbashi) three days per week
技術スタック
必須スキル
5+ years in software engineering with strong data structures, algorithms, design patterns and best practices
2+ years UNIX-based systems (Linux), Python, and PyTorch/TensorFlow
2+ years full MLOps lifecycle: data cleansing/sampling/curation, preprocessing, distributed training, evaluation, deployment, inference optimization (cloud and edge)
Docker and CI systems (GitHub Actions)
Business-level English (technical writing)
歓迎スキル（該当する場合）
2+ years with Apache Spark, Airflow, Flyte, Flink, Ray or similar ML pipelines
2+ years with Rust and/or C++, Bazel, systems-level debugging
SIMD/SIMT, GPU programming, multithreading
Terraform, AWS, Observability, Kubernetes in production
BigQuery, Snowflake, or AWS Redshift in production
Experience in self-driving, robotics, computer vision, or motion planning