Principal Controls Engineer

Foundation•San Francisco, CA

10d•Onsite

About The Position

To develop and optimize the control systems that enable precise, smooth, and safe motion in our humanoid robots. To bridge the gap between AI-generated action models (e.g., RL policies) and the robot’s physical execution by designing robust motion planning, state estimation, and lower-level control architectures. To create full-body controllers that seamlessly integrate inputs from action models with real-time feedback, ensuring high performance in complex environments. To ensure our robots can execute tasks reliably, from manipulation to locomotion, while maintaining stability and efficiency. To scale our controls stack for deployment across various robot configurations, supporting a wide range of customer applications.

Requirements

Strong background in control theory, including linear and nonlinear systems, feedback control, stability analysis, and optimal control methods.
Proficiency in state estimation techniques, such as Kalman filters, particle filters, or sensor fusion for estimating robot pose, velocity, and dynamics from noisy sensor data (e.g., IMUs, encoders, vision).
Experience with low-level control strategies, including PID controllers, torque control, velocity control, and position control for actuators and joints.
Skilled in developing full-body controllers that integrate multiple degrees of freedom (DoF), handling kinematics, dynamics, and inverse kinematics.
Ability to translate high-level action outputs (e.g., from AI or RL models) into low-level commands such as joint angles, torques, velocities, and forces.
Familiarity with optimization techniques for trajectory planning, model predictive control (MPC), and real-time feedback integration.
Proficiency in programming languages such as C++ and Python.
Experience with robotics middleware (e.g., ROS2) and real-time control systems.
Strong understanding of numerical methods and computational algorithms for control.
Ability to analyze complex motion requirements and design control algorithms to meet them.
Comfortable troubleshooting control issues in a physical robot and iterating quickly to improve performance.
Comfortable working with hardware-in-the-loop systems and integrating controls with physical robots.
Collaborates effectively with AI researchers, mechanical engineers, and electrical engineers to ensure seamless system integration.
Minimum of 5 years of experience in controls engineering, motion planning, or a similar field; experience from humanoids is ideal but not required. Transferable skills from drones, autonomous vehicles, or industrial robotics are welcome if they include state estimation and low-level controls.
Bachelor’s or Master’s degree in Robotics, Mechanical Engineering, Computer Science, or a related field.

Nice To Haves

Experience with reinforcement learning (RL) for control policies is a strong plus.
A PhD is a plus.
Experience with real-time operating systems and control frameworks is highly desirable.