Radiant Robotics is advancing intelligent automation and robotics through magnetic levitation controls, and this role sits at the intersection of controls systems and machine learning. In San Leandro, CA (onsite), you will work on simulation and real hardware to design, validate, and deploy control strategies for dynamic magnetic systems.
What you’ll do
- Design control strategies including PID, model predictive control, and learned policies, then recommend which approaches to deploy.
- Develop and optimize control algorithms using machine learning methods such as reinforcement learning and predictive modeling.
- Design, develop, and deploy machine learning models where they add value, including disturbance modeling and trajectory optimization.
- Build and maintain simulation models covering dynamic magnetic forces, moving payloads, and the sensing chain, and validate them against bench measurements.
- Run comparative experiments to build and validate simulations of different control strategies.
- Define and track performance metrics for the magnetic levitation system, such as positional error, and report results against targets.
- Document models, assumptions, and results so other engineers can analyze, reproduce, and extend the work.
- Partner with firmware and electronics engineers to move proven strategies from simulation to real hardware and tune them in the loop.
What you bring
- Master’s degree or higher in electrical engineering, mechanical engineering, robotics, computer science, applied mathematics, or a related field.
- Working knowledge of control theory, including PID design and tuning, state-space methods, and model predictive control (MPC) (including deriving a plant model, linearizing, and explaining tradeoffs).
- Applied machine learning experience, including at least one project where a learned model or policy improved a physical or simulated system.
- Experience building dynamic simulations of physical systems and comparing them to real measurements.
- Proficiency in Python, C++, or MATLAB/Simulink for controls and machine learning development.
- Clear written and verbal communication, including the ability to explain controller behavior to engineers outside the controls discipline.
- Comfort using AI coding tools such as Claude Code or similar for simulation development.
Tools and technologies
- PID, model predictive control (MPC)
- reinforcement learning
- Python, C++, MATLAB/Simulink
- Claude Code, FEMM, COMSOL
Compensation and benefits
- $130,000 - $180,000 per year (USD)
- Paid time off
- 401(k)
- Competitive salary
- Opportunity for advancement
- Stock options plan
Preferred qualifications
- Experience deploying controllers on embedded targets such as microcontrollers or real-time Linux, including hardware-in-the-loop tuning.
- Experience with multi-physics or finite element modeling of magnetic fields, for example with FEMM or COMSOL.
- Experience designing, analyzing, or fabricating systems that generate or manipulate magnetic fields.