Physics Simulation Intern

Join NVIDIA's Physics Simulation Team as an intern and work on a feature within Newton physics or build a user workflow using Newton to solve customer needs and accelerate robot learning. Possible areas include physics calibration and validation, simulation fidelity and performance, and integration into robotics training and evaluation workflows.

Responsibilities

• Collaborate with senior robotics and simulation engineers to understand customer needs and develop a Newton core feature or reusable workflow.
• Take ownership of a scoped contribution from investigation through implementation and evaluation, with guidance and code review from experienced engineers.
• Design experiments to demonstrate measurable improvements in correctness, physical fidelity, reliability, performance, or usability.
• Use AI-assisted development tools to explore approaches, implement and debug code, and validate results.
• Contribute tests, documentation, and maintainable software to help users get reliable simulation results and accelerate robot learning.

Requirements

• Pursuing an MS or PhD in Computer Science, Robotics, Physics, or a related field.
• Strong analytical abilities, with foundations in mathematical analysis, linear algebra, numerical algorithms, and physics.
• Strong programming skills and ability to understand, modify, and debug existing codebases.
• Proficiency with AI-assisted development and coding agents.
• Good judgment in choosing meaningful problems and defining manageable experiments.
• Solid software engineering foundations, clear technical communication, and willingness to take responsibility for project quality.

Nice to Have

• Experience with physics engines like Newton, PhysX, or MuJoCo.
• Hands-on experience in robotics, control, robotic manipulation, or machine learning for robotics.
• Experience with NVIDIA's robotics software stack, such as Isaac Sim and Isaac Lab.
• Experience with GPU programming and performance analysis using NVIDIA Warp or CUDA.
• A project demonstrating careful experimental design, physical or numerical validation, and reusable software or research results.