about 5 hours ago
Responsibilities
- Implement and deploy reinforcement learning algorithms for dynamic locomotion and manipulation tasks on physical humanoid robot hardware.
- Drive development from simulation prototyping through policy transfer and fine-tuning on the robot.
- Optimize and scale reinforcement learning training pipelines and contribute to high-throughput simulation and distributed training infrastructure.
- Mentor junior engineers through technical guidance, code reviews, and reinforcement learning best-practice sharing.
- Collaborate with robotics and hardware teams to diagnose system-level issues and enable more complex learned behaviors.
- Analyze and present hardware results to guide technical direction and demonstrate progress.
- Develop and refine motion retargeting pipelines that convert mocap and teleoperation demonstrations into reference trajectories for reinforcement learning.
Requirements
- Deep hands-on expertise with reinforcement learning frameworks such as PyTorch and JAX and high-fidelity physics simulators such as MuJoCo and IsaacGym.
- Mastery of Python and strong proficiency in C++.
- Experience building or using large-scale distributed training pipelines and optimizing them for rapid iteration.
- Strong theoretical knowledge of reinforcement learning, including imitation learning, model-based reinforcement learning, and sim-to-real transfer.
- Strong understanding of robot dynamics and controls theory.
- PhD or MS in Computer Science, Robotics, or a related field.
- A proven record of deploying learning-based policies on physical robotic systems, especially legged robots or manipulators.
- Experience mentoring or providing technical guidance to other engineers.
- A strong publication record in relevant conferences or journals such as CoRL, RSS, or ICRA is a significant plus.
Benefits
- Direct-hire position.
About Apptronik
Apptronik is a human-centered robotics company developing AI-powered robots to support humanity in every facet of life. Our humanoid robot, Apollo, is designed to collaborate thoughtfully with humans—initially in critical industries such as manufacturing and logistics, with future applications in healthcare, the home, and beyond. Apollo is the culmination of nearly a decade of development, drawing on Apptronik’s extensive work on 15 previous robots, including NASA’s Valkyrie robot. Apptronik started out of the Human Centered Robotics Lab at the University of Texas at Austin and has more than 350 employees. Learn more at apptronik.com