Teams at Physical Intelligence

Own and optimize infrastructure for large-scale ML training and inference. Build JAX-based pipelines, manage GPU and TPU workloads, improve distributed performance, and create systems for scheduling, monitoring, checkpointing, debugging, and reproducible experimentation. Partner with researchers to translate modeling ideas into production training runs and contribute to core training code supporting new architectures, modalities, and evaluation metrics.

This role involves executing cross-functional projects, driving partnerships, managing deployment efforts, and supporting communications with investors and external parties.

Design and implement control algorithms for robotics, develop real-time control loops, and debug complex hardware/software issues while collaborating with cross-functional teams.