Robotics Machine Learning Engineer - Platforms for Vision-Language Action Foundation Models

This role focuses on building and maintaining the infrastructure enabling robotics foundation models for dexterous manipulation.
You will develop hardware platforms, sensors, interfaces, data logging, and controls to support researchers.
You will design scalable pipelines for heterogeneous multimodal data (images, text, video, touch, depth, proprioception) and ensure efficient data handling.
You will implement high-throughput inference, low-latency model serving, and robust policy–environment communication for both simulation and real robots.
You will create tooling for experiment tracking, evaluation, visualization, and leaderboard management using systems like Weights & Biases, MLflow, or ClearML.
The role requires collaborating with researchers and contributing to publications while iterating rapidly across simulated and real-world robotics experiments.