Robotics Real-Time Control Systems Engineer

OpenAI; San Francisco (CA); On-site; USD 380,000 - 460,000; Full time

Job summary

OpenAI is seeking a Control Systems Software Engineer, Robotics in San Francisco to architect, implement, tune, and verify real-time control infrastructure for intelligent robot behavior. This hands-on role focuses on actuation, dynamics, and hardware interaction onsite five days a week.

You will collaborate across mechanical, electrical, and software teams, work with ML engineers on reinforcement learning models, and help shape whole-robot performance from low-level actuators up to high-level

Qualifications

• Strong robotics background with manipulators, mobile robots, or similar systems.
• Experience with actuation, control systems, dynamics, estimation, or whole-body motion.
• Hands-on with robotic hardware, including commissioning, debugging, testing, and iteration.

Responsibilities

• Design and implement real-time control algorithms for robotic systems, including motion control and state estimation.
• Define the control architecture from actuators to whole-robot behavior.
• Characterize parameters through experimentation, testing, and verification.
• Work with ML engineers to implement reinforcement learning models.
• Collaborate across mechanical, electrical, and software teams to integrate control with sensing and actuation.
• Inform mechanical/electrical design to maximize capability and flexibility.
• Create the control system architecture and abstraction levels from actuators to policy.

Skills

Real-time control

C++

Rust

Robotics background

Hardware interaction

Job description

OpenAI is seeking a Control Systems Software Engineer, Robotics in San Francisco to architect, implement, tune, and verify real-time control infrastructure for intelligent robot behavior. This hands-on role focuses on actuation, dynamics, and hardware interaction onsite five days a week.

You will collaborate across mechanical, electrical, and software teams, work with ML engineers on reinforcement learning models, and help shape whole-robot performance from low-level actuators up to high-level