About the Role
NVIDIA Research’s AI-Mediated Reality and Interaction team is seeking a research scientist. Our mission is to create interactive physical AIs capable of operating in dynamic 4D worlds. We conduct fundamental research on the algorithms, theory, and application of human-AI interaction and 4D world modeling using new ideas in artificial intelligence (AI), neural rendering, graphics, generative modeling, large language models, human behavior understanding, and digital human creation. This research impacts many existing NVIDIA products, including Isaac, Omniverse, Metropolis, NuRec, XR, AI4Media, and helps create new ones. We are looking for a highly creative, independent, and self-starting research scientist to help us invent this future. Join a diverse research group that collaboratively works on challenging and meaningful interdisciplinary problems with passion, drive, and curiosity; that consistently publishes and demonstrates cutting-edge technology at top venues in computer vision, artificial intelligence, and graphics; and that values real impact on NVIDIA products and the industry at large.
What you’ll be doing:
- Propose, research, prototype, and test innovative research ideas.
- Publish at top conferences and patent research ideas.
- Collaborate with other research team members, external researchers, and mentor interns.
- Create working interactive demonstrations to showcase your research ideas.
- Participate in technology transfer with engineers around NVIDIA as ideas "graduate" from research to product.
What we need to see:
- Ph.D. in Computer Science/Electrical Engineering or equivalent research experience.
- Outstanding research track record in top-tier conferences (CVPR, ICCV, ECCV, SIGGRAPH, NeurIPS, ICLR, etc.).
- Proficiency in theory and practice of AI and computer vision algorithms.
- Experience with 3D graphics development.
- Strong programming skills in Python, C++, CUDA.
- Proficiency with distributed deep learning training frameworks, e.g., PyTorch.
- Excellent communication skills.