Research Scientist, Physical AI
NVIDIA
About the Role
Join our groundbreaking research team as we revolutionize the future of Physical AI through innovative generative models. As a Research Scientist specializing in Generative AI for Physical AI, you will be at the forefront of developing next-generation algorithms that bridge the gap between virtual and physical realms, working with state-of-the-art technology and massive computational resources.
Key Responsibilities
- Pioneer revolutionary generative AI algorithms for physical AI applications, focusing on advanced video generative models and video-language models.
- Architect and implement sophisticated data processing pipelines for training data in Generative AI and Physical AI systems.
- Design and develop cutting-edge physics simulation algorithms to enhance Physical AI training.
- Scale and optimize large-scale training systems to efficiently harness the power of 20,000+ GPUs for training foundation models.
- Author influential research papers to share discoveries with the global AI community.
- Drive innovation through close collaboration with research teams, internal product groups, and external researchers.
- Facilitate technology transfer and contribute to open-source initiatives.
Requirements & Tech Stack
- PhD in Computer Science, Computer Engineering, Electrical Engineering, or a related field (or equivalent experience).
- Deep expertise in PyTorch and related libraries for Generative AI and Physical AI development.
- Strong foundation in diffusion models, vision language models, and reasoning models and their applications.
- Proven experience with reinforcement learning algorithms and implementations.
- Robust knowledge of physics simulation and its integration with AI systems.
- Demonstrated proficiency in 3D generative models and their applications.
Qualifications (Ways to Stand Out)
- Publications or contributions to major AI conferences (e.g., ICLR, NeurIPS, ICML, CVPR, ECCV, SIGGRAPH, ICCV).
- Experience with large-scale distributed training systems.
- Background in robotics or physical systems.
- Open-source contributions to prominent AI projects.
- History of successful research-to-product transitions.