About the job
Do you have a passion for applying state-of-the-art controls and machine learning techniques to solve real-world problems? As a Robotics Engineer on the Spot Manipulation team, you will be a key contributor to tackling some of the most challenging technical problems in robotics. Your role will focus on the hands-on development and implementation of a data-driven manipulation stack. We are investing in reinforcement learning and behavioral cloning as key technologies for achieving dexterous and robust loco-manipulation that can be deployed in real-world environments. Success in this position requires ambitious ideas, rapid and meticulous execution, and a collaborative mindset within a dynamic, fast-paced team.
Responsibilities
Design and implement novel systems to enable next-gen loco-manipulation behaviors.
Rapidly iterate on manipulation prototypes to move from concept to on-robot tests.
Execute rigorous testing and debugging of designs directly on our in-house robot fleet.
Collaborate closely with other project teams to build a cross-embodiment data flywheel.
Qualifications
Minimum
MS with 3+ years of industry experience or PhD in Robotics, Machine Learning, Computer Science, or related field.
Experience training and deploying RL policies for complex behaviors on real robots or simulated characters
Experience integrating perception in learned behaviors.
Proficiency in Python & C++ programming language.
Strong analytical and debugging skills in a hardware-centric environment.
Preferred
Experience with legged loco-manipulation.
Strong background in behavior cloning and curating large demonstration datasets.
Ideas to close the embodiment gap between human/humanoid data and quadrupeds.
Previous experience in high-growth or fast-paced robotics environments.