Robotics Autonomy Engineer - Locomotion

Job description

[FieldAI](http://www.fieldai.com)’s Irvine team is where embodied AI meets real robots, real sensors, and real field deployments. Based in the heart of Southern California’s robotics ecosystem, we build risk-aware, reliable, field-ready AI systems that solve the hardest problems in robotics and unlock the full potential of embodied intelligence. If you want your work to ship, get tested on hardware, and improve through real deployments, Irvine is the place. We go beyond typical data-driven approaches or pure transformer-only architectures, combining rigorous engineering with learning systems proven in globally deployed solutions that deliver results today and get better every time our robots run in the field. **About the Job** Field AI is building the future of autonomy—from rugged terrain to real-world deployment. We’re on a mission to develop intelligent, adaptable robotic systems that operate beyond simulation and thrive in unpredictable environments. As our Robotics Autonomy Engineer – Locomotion, you’ll lead the development and deployment of state-of-the-art controllers for legged and humanoid robots. You’ll be part of a deeply technical team advancing real-world robotic capabilities through cutting-edge research, simulation tools, and field validation. If designing locomotion systems that can navigate complex, dynamic environments excites you, and you want to work where your code hits the ground (literally)—this is your role. This is Field AI. ### What You’ll Get To Do - Architect and implement scalable reinforcement learning (RL) pipelines for locomotion and whole-body control on legged and humanoid robots - Design policy architectures, training curricula, and domain randomization strategies that close the sim-to-real gap - Integrate GPU-accelerated, physics-based simulation environments with custom, distributed training workflows - Train locomotion policies from human motion data using imitation learning and motion retargeting, and distill them into compact policies that run on the robot - Create agile, robust, and terrain-aware (perceptive) locomotion behaviors for quadruped and humanoid platforms, and validate them on real hardware - Solve real-world challenges in balance and push recovery, contact-rich dynamics, high degree of freedom (DOF) whole-body coordination, and terrain variability - Automate evaluation across domain-randomized scenarios and maintain simulation infrastructure that enables rapid prototyping, validation, and reproducibility - Work closely with systems engineers, perception experts, and embedded teams, incorporating real-world telemetry and field data to continuously improve generalization - Lead deployment workflows from experiment through lab testing to field robot validation ### What You Have - Master’s degree or higher in Robotics, Computer Science, Engineering, or related field (PhD strongly preferred) - Deep expertise in reinforcement learning for continuous control - 2+ years of experience developing and deploying locomotion policies on real robotic systems (preferred) - Hands-on experience with legged robot platforms (quadrupeds, bipedal/humanoid systems, or exoskeletons) - Proficiency with simulation tools such as Isaac Gym, Isaac Lab, MuJoCo, or PyBullet - Strong command of sim-to-real transfer and a track record of bridging the gap successfully - Solid understanding of contact dynamics, control theory, and kinematics - Strong Python and/or C++ development skills in Linux-based development environments - Familiarity with machine learning frameworks (PyTorch, JAX, TensorFlow) - A passion for building things that move in the real world ### The Extras That Set You Apart - 3+ years of experience in an industry or startup robotics setting - Experience optimizing and deploying learning based controllers on resource constrained robotic platforms (ONNX Runtime, NVIDIA TensorRT, real time onboard inference) - Publications or open-source contributions in locomotion, reinforcement learning, or control (e.g., CoRL, RSS, ICRA, IROS) - Familiarity with ROS/ROS2 or custom middleware for real-time control - Background in manipulation, loco-manipulation, or whole-body coordination - Experience combining learned policies with model predictive control (MPC) or whole-body controllers - Experience with motion imitation pipelines from motion capture or teleoperation data - Experience debugging sim-to-real issues at scale - Contributions to reinforcement learning libraries or simulation platforms - Prior work on multi-agent learning or terrain-adaptive control systems