Robotics AI Foundation Models Intern: Training & Inference

Roboticsconnected; Santa Clara (CA); On-site; USD 34,000 - 55,000; Part time

Job summary

NVIDIA’s DevTech team is seeking a passionate intern to work on robotics foundation model training and inference, powering intelligent robotic systems on NVIDIA platforms. You will develop training pipelines, optimize inference performance, and apply modern DL/ML techniques to real-world robotics challenges.

Ideal candidates are pursuing a Master’s or PhD in CS/Robotics/AI, with strong Python, C/C++, and CUDA skills, plus experience in PyTorch and transformer architectures.

Qualifications

• Pursuing a Master’s or PhD in Computer Science, Robotics, AI/ML, Electrical Engineering, or related field.
• Strong programming skills in Python; familiarity with C/C++ and CUDA a plus.
• Experience with deep learning frameworks (PyTorch).
• Understanding of transformer architectures and large-scale model training.
• Basic knowledge of robotics concepts (kinematics, control, perception, or simulation).
• Curious, self-motivated, and excited about solving open-ended challenges at NVIDIA.

Responsibilities

• Develop and optimize training pipelines for robotics foundation models on NVIDIA GPU clusters.
• Profile, analyze, and optimize model inference performance on NVIDIA platforms.
• Apply modern DL/ML techniques to real-world robotics problems.
• Collaborate with NVIDIA teams to influence design of architectures, software platforms, and programming models.
• Build samples, reference implementations, and documentation to educate developers on NVIDIA robotics AI technologies.
• Discuss approach and results with senior DevTech engineers.

Skills

Python

C/C++

CUDA

PyTorch

Transformer architectures

Robotics concepts

Education

Master’s or PhD in CS/Robotics/AI/ML

Tools

DeepSpeed

Megatron-LM

FSDP

NeMo

CUDA toolkit

Job description

NVIDIA’s DevTech team is seeking a passionate intern to work on robotics foundation model training and inference, powering intelligent robotic systems on NVIDIA platforms. You will develop training pipelines, optimize inference performance, and apply modern DL/ML techniques to real-world robotics challenges.

Ideal candidates are pursuing a Master’s or PhD in CS/Robotics/AI, with strong Python, C/C++, and CUDA skills, plus experience in PyTorch and transformer architectures.