Humanoid Research Scientist and Engineer

Reward AI; San Francisco (CA); On-site; USD 120,000 - 180,000; Full time

Job summary

Reward AI is building frontier intelligence for robots, connecting learning, control, and hardware to enable capable, reliable behavior across varied bodies and tasks. The role spans research and implementation, offering exposure to simulation and real hardware work, from initial experiments to repeatable robotic behaviors in dynamic environments.

Join a multidisciplinary team tackling locomotion, dexterous manipulation, and real-time control, with opportunities to experiment in both simulation

Qualifications

• Background in robotics, controls, ML, or mechanical engineering through research or practical work.
• Experience in robot learning, dynamics, motion planning, or state estimation.
• Ability to develop and debug robotics software in Python or C++, with sim or DL tools.
• Hands-on work with humanoids, legged robots, or complex robotic systems.
• Proven ability to build and evaluate working robotic systems and solve cross-domain problems.

Responsibilities

• Develop methods for whole-body control, locomotion, dexterous manipulation, and coordinated movement.
• Build training and evaluation workflows using simulation, human motion, and data sources.
• Investigate transfer from simulation to hardware and robustness to changes in contact, objects, terrain, dynamics.
• Integrate learned behaviors with perception, state estimation, planning, and real-time control systems.
• Run hardware experiments with safeguards, diagnose failures, improve robot capability and reliability.

Skills

Robotics background

Controls

Machine learning

Mechanical engineering

Research experience

Python

C++

Simulation tools

Deep learning tools

Tools

Python

C++

Simulation

Deep learning

Job description

At Reward AI, we are building frontier intelligence for any robot. Our work brings together machine learning, robotics, and hardware to help robots learn from people and perform useful work across bodies, tasks, and environments. We welcome researchers, engineers, and people with experience in operations and product from a wide range of backgrounds.

Help humanoid robots move and work with greater capability, adaptability, and reliability. You will connect learning, control, and physical systems to develop behaviors that combine locomotion and manipulation. The role spans research and implementation, with opportunities to work in simulation and on real robots across the path from an initial experiment to repeatable behavior.

Responsibilities

• Develop methods for whole-body control, locomotion, dexterous manipulation, and coordinated movement.
• Build training and evaluation workflows using simulation, human motion, and other relevant data sources.
• Investigate transfer from simulation to hardware and improve robustness to changes in contact, objects, terrain, and robot dynamics.
• Integrate learned behaviors with perception, state estimation, planning, and real-time control systems.
• Run hardware experiments with appropriate safeguards, diagnose failures, and work with the team to improve robot capability and reliability.

Preferred qualifications

• A background in robotics, controls, machine learning, mechanical engineering, or a related field through research or practical experience.
• Experience in one or more of robot learning, dynamics, motion planning, optimal control, or state estimation.
• Ability to develop and debug robotics software in Python or C++, with familiarity with simulation or deep learning tools.
• Hands-on work with humanoids, legged robots, manipulators, or other complex robotic systems.
• Evidence of building and evaluating working systems, sound experimental judgment, and an interest in solving problems across software and hardware.