Real-Time Robotics Deployment Engineer

physicalintelligence; San Francisco (CA); On-site; USD 150,000 - 210,000; Full time

Job summary

Physical Intelligence in San Francisco seeks a Robotics Software Engineer to design on-site control systems that integrate AI models with robots operating on customer floors. You will build teleoperation interfaces, ensure low-latency performance, and contribute to reliable, observable software.

You will collaborate with researchers, platform engineers, and robot operators to fine‑tune latency, throughput, and real‑time correctness across the full stack, from compute to video pipelines, enabling

Qualifications

• Strong programming in C++, Rust, or Python.
• Experience with real-time or near real-time systems.
• Ability to optimize across the entire stack—drivers, networking, compute, video frameworks, and distributed components.

Responsibilities

• Design remote teleoperation interfaces for robots with many degrees of freedom.
• Optimize for low-latency network performance to handle weak internet areas.
• Profile and optimize compute, I/O, memory, scheduling, networking, and storage for real-time constraints.
• Develop and optimize real-time video streaming systems where timing matters.
• Build tooling for profiling, tracing, and debugging timing issues across distributed systems.
• Collaborate with researchers, hardware engineers, and operators to integrate optimized pipelines into production workflows.

Skills

C++/Rust/Python

Real-time systems

Video pipelines

Profiling tools

Determinism/throughput

Cross-functional collaboration

Tools

Perf

eBPF

GPU profilers

Job description

Physical Intelligence in San Francisco seeks a Robotics Software Engineer to design on-site control systems that integrate AI models with robots operating on customer floors. You will build teleoperation interfaces, ensure low-latency performance, and contribute to reliable, observable software.

You will collaborate with researchers, platform engineers, and robot operators to fine‑tune latency, throughput, and real‑time correctness across the full stack, from compute to video pipelines, enabling