AI Robotics Simulation Intern

NIO • San Jose, CA

1d • $38 - $46

About The Position

About NIO NIO is a pioneer and a leading company in the premium smart electric vehicle market. Founded in November 2014, NIO’s mission is to shape a joyful lifestyle. NIO aims to build a community starting with smart electric vehicles to share joy and grow together with users. NIO designs, develops, jointly manufactures and sells premium smart electric vehicles, driving innovations in next-generation technologies in autonomous driving, digital technologies, electric powertrains and batteries. NIO differentiates itself through its continuous technological breakthroughs and innovations, such as its industry-leading battery swapping technologies, Battery as a Service, or BaaS, as well as its proprietary autonomous driving technologies and Autonomous Driving as a Service, or ADaaS. NIO’s product portfolio consists of the ES8, a six-seater smart electric flagship SUV, the ES7 (or the EL7), a mid-large five-seater smart electric SUV, the ES6, a five-seater all-round smart electric SUV, the EC7, a five-seater smart electric flagship coupe SUV, the EC6, a five-seater smart electric coupe SUV, the ET7, a smart electric flagship sedan, and the ET5, a mid-size smart electric sedan. We are building next-generation dexterous manipulation intelligence for embodied robotics systems. Our work spans contact-rich manipulation, physics-based simulation, and scalable data generation for robotic learning systems. This internship will focus on advancing our high-fidelity simulation infrastructure to support contact-rich robotic manipulation. Project Scope The intern will contribute to one or more of the following areas: High-Fidelity Contact Simulation Improve geometric modeling and mesh processing pipelines for robotic hands and objects. Develop robust surface reconstruction and mesh conditioning tools for simulation assets. Analyze mesh quality, collision stability, and contact robustness. Physics-Driven Simulation Infrastructure Design automated pipelines for physics parameter identification (System ID) to calibrate contact dynamics (e.g., stiffness, damping, friction profiles). Develop tools for systematic sensitivity analysis and domain randomization of simulation parameters. Build robust simulation wrappers and configuration modules to manage contact-rich environments across different backend solvers (e.g., MuJoCo, Isaac). Rendering & Visualization for Simulation Debugging Build real-time, high-performance visualization tools for contact points, contact forces, friction cones, and constraint violations. Develop zero-copy, tensor-native GPU debugging overlays to inspect massively parallel simulations without bottlenecking data generation throughput. Design intuitive UI/UX for robotics researchers to toggle and filter complex contact interactions during live policy rollouts. Simulation-to-Real Gap Analysis Design controlled experiments to measure contact dynamics fidelity by comparing simulation trajectories against real-world hardware logs. Evaluate simulation robustness under sim-to-real transfer conditions: Contact state perturbations and sensor noise Imperfect object meshes (scanned vs. ground truth) Physical parameter domain randonmization Produce quantitative reports to guide the calibration of our physical simulation infrastructure.

Requirements

• PhD or strong MS student in Computer Science, Robotics, Computer Graphics, or related field.
• Strong C++ and Python programming skills.
• Solid foundation in 3D geometry processing, mesh generation, or surface reconstruction.
• Experience with rendering APIs (OpenGL/WebGL/Vulkan) and GPU programming.
• Strong debugging skills and system-level thinking.

Nice To Haves

• Hands-on experience with robotics simulators (MuJoCo, Isaac, Bullet, etc.).
• Knowledge of collision detection algorithms (e.g., GJK, EPA) and contact modeling (LCP, soft contacts).
• Experience with CUDA, parallel computing, or tensor-native operations (e.g., PyTorch).
• Familiarity with writing Python bindings for C++ code (e.g., pybind11) and integrating them into ML pipelines.
• Prior experience working with real robot hardware or sim-to-real transfer pipelines.

Responsibilities

• Massively parallel GPU simulation architecture and scalable infrastructure for robot learning.
• Advanced contact mechanics and numerical methods for contact-rich robotics.
• System Identification (System ID) and practical sim-to-real transfer techniques.
• Industry-scale research execution for embodied foundation models.
• Production-quality, well-documented simulation tools, wrappers, or mesh processing modules integrated into our codebase.
• A comprehensive technical report documenting: Experimental design (e.g., for System ID or stress-testing).
• Evaluation methodology for contact stability and the sim-to-real gap.
• Quantitative benchmarking results.
• An internal presentation to the robotics team demonstrating the new tools during live policy rollouts.

What This Job Offers

• Job Type: Full-time
• Career Level: Intern