Research Scientist Intern: Differentiable Robotics

Meta; Redmond (WA); On-site; USD 34,000 - 55,000; Part time

Job summary

Meta Reality Labs Research is seeking an intern to advance next-generation physics simulators for robotics. The role focuses on differentiable simulation of contact mechanics and training dexterous manipulation policies, with integration into a broader software stack.

The internship lasts 12–24 weeks with multiple start dates, requiring a PhD candidate and strong C++/Python skills, CUDA, and PyTorch experience.

Qualifications

• Pursuing a PhD in computer science, computer graphics, computational mechanics, machine learning, computational mathematics or a relevant technical field.
• Research skills involving defining problems, exploring solutions, and analyzing and presenting results.
• Experience implementing numerical methods and/or deep learning models in a large-scale codebase.
• Proficient in C++ and/or Python.
• Must obtain work authorization in the country of employment and maintain ongoing work authorization; proven track record of significant results (grants, fellowships, patents, or first-authored publications).
• Experience with high performance/parallel computing (e.g., CUDA).
• Experience with training deep learning models and familiarity with PyTorch.
• Experience working and communicating cross-functionally in a team environment.
• Intent to return to degree program after completion of the internship.
• Availability for minimum 16 consecutive week internship.

Responsibilities

• Develop, implement, and test novel numerical methods for differentiable simulation of contact mechanics.
• Develop and implement strategies for training dexterous robot manipulation policies that are efficient and robust under the complexity of contact.
• Implement and optimize the methods devised in a larger software infrastructure, and collaborate with research scientists and software engineers to build demos and experiences based upon the developed tools.

Job description

Meta Reality Labs Research is seeking an intern to advance next-generation physics simulators for robotics. The role focuses on differentiable simulation of contact mechanics and training dexterous manipulation policies, with integration into a broader software stack.

The internship lasts 12–24 weeks with multiple start dates, requiring a PhD candidate and strong C++/Python skills, CUDA, and PyTorch experience.