AI Research Scientist, Large Scale Behavior Cloning

As an AI Research Scientist, Large Scale Behavior Cloning on the Atlas team, you will drive the development of scalable, generalizable behavior models that enable our humanoid robot to master complex mobile manipulation tasks and generalize to new situations. You will work alongside a team of leading experts in robotics, machine learning, and control, contributing to groundbreaking advancements in robot autonomy and intelligence.

About the Role:

You will play a key role in developing and deploying algorithms that empower Atlas to learn from experience, generalize to new environments, and perform a wide range of manipulation tasks. You will be involved in designing and implementing solutions that address the core challenges of robot learning, to train generative models on large-scale demonstration data leveraging multiple data sources, including teleoperation, human demonstrations, and internet data.

How you will make an impact:

• Implement large-scale imitation learning pipelines to train powerful generalist agents.

• Design hierarchical planning and control systems that leverage learned skills and knowledge.

• Integrate learning-based modules into complex robotic systems to enable broadly capable manipulation.

• Develop and implement algorithms that enable robots to efficiently explore and learn in the real world.

• Write high-quality (documented, tested, maintainable) C++ and Python code.

We are looking for:

• MS with 3+ years of industry experience or PhD in Robotics, Computer Science, Machine Learning, or a related field

• Prior experience training and deploying learned policies for complex behaviors in robots or simulated characters.

• Experience with common ML tools, architectures, and workflows.

• Strong analytical and debugging skills.

Nice to have:

• PhD in Robotics, Computer Science, Machine Learning, or a related field

• Experience with modern behavior cloning technologies.

• Experience with robot hardware and real-world robot deployments.

#LI-JM1