Physical Intelligence

ML Infra Engineer responsible for designing and building a scheduling and compute layer for large-scale AI model training across heterogeneous GPU/TPU clusters. This role focuses on intelligent resource allocation, utilization, fault tolerance, and making distributed training seamless, extending to inference and robot deployment.

This role involves teleoperating robotic arms to collect high-quality demonstration data for training foundation models for physical world AI. The operator will perform various tasks, maintain data quality, and meet collection metrics. It's a hands-on, metrics-driven position focused on generating training data for AI-powered robots.

ML Infra Engineer focused on scaling and optimizing training systems and core model code, managing GPU/TPU compute, job orchestration, and building efficient JAX training pipelines. Collaborates with researchers to translate ideas into production training runs.

ML Infra Engineer to scale and optimize training systems and core model code, managing GPU/TPU compute, job orchestration, and JAX training pipelines. Collaborates with researchers to translate ideas into production training runs.

Controls Engineer for Physical Intelligence, designing and implementing algorithms for robot control, including PID, LQR, MPC, and neural-network-driven control. Responsibilities include building and validating models, developing real-time loops, owning robotic bring-up, debugging complex systems, and working with sensor/actuator subsystems. Requires strong understanding of model-based control, Python/C++ proficiency, and real-time loop tuning skills. Bonus for robot learning or manipulation experience.