Senior Machine Learning Engineer

This role is part of Newmo’s Autonomous Driving Development Office and supports the in-house development of newmo Autonomy, an autonomous-driving system for taxi operations. The position develops and implements learning-based end-to-end autonomous-driving models for Physical AI, as an initial core-member role in the project.

The models receive sensor inputs such as camera images and LiDAR point clouds and perform route planning and vehicle control. The approach differs from a rule-based modular system that separates recognition, prediction, and planning, aiming instead to connect sensor input directly to vehicle control through an end-to-end model.

Responsibilities include developing these autonomous-driving models, conducting offline evaluations, and evaluating model performance with a simulator. The role also establishes a cycle for improving the models based on evaluation results. Additional work includes optimizing models for edge devices and assessing their performance using real vehicles.

The development effort targets the commercial deployment of Level 4 autonomous driving in 2028. The work applies machine-learning and foundation-model technologies to complex real-world driving problems in the Physical AI and VLA areas. Simulation and vehicle testing are both used to evaluate the models. Training is performed at scale with GPU servers and large volumes of sensor data collected by the company, including data from its own fleet and driving and safety data.

Named technical areas include machine learning, end-to-end autonomous-driving models, Physical AI, VLA, foundation models, camera images, LiDAR point clouds, simulators, edge-device optimization, GPU servers, and sensor data. The listed work location is Autonomy Garage Tokyo in Ota, Tokyo, Japan.