ML Infra Engineer, Compute
Physical Intelligence · San Francisco
Roboticsonsite
At a glance
Highlights
- Own intelligent scheduling across thousands of accelerators for robotics AI.
- Build placement logic that routes training jobs automatically by cost and priority.
- Extend orchestration to inference, deploying models to edge devices on robots.
Ideal candidate
Systems-oriented engineer who loves resource allocation and fault tolerance and wants to own scheduling end-to-end. You thrive unblocking researchers and making large-scale distributed training seamless and efficient.
Requirements
- Strong software engineering fundamentals
- Experience building or operating job scheduling / resource management systems at scale
- Experience with large-scale compute clusters (GPU and/or TPU)
- Familiarity with schedulers and orchestration systems (SLURM, Kubernetes, GKE, K3S, or internal equivalents)
- Comfort reasoning about resource allocation, bin-packing, priority scheduling, and multi-tenancy
- Understanding of how ML training workloads behave — long-running, multi-node, sensitive to stragglers, topology-dependent
- Bias toward owning systems end-to-end, from design to operation
- Enjoy working closely with researchers and unblocking fast-moving projects
Lab profile
About Physical Intelligence
Developing general-purpose AI foundation models to power robots for any physical task.
What they build
- General-purpose robot foundation models
- Learning algorithms for robot control
- Robotic systems and platforms
Known for
- Foundation model and system-building research
- Large-scale multimodal robot data operations
- Real-world fleet deployment and operations
- Prototype-to-production hardware engineering
Posted · Last seen