As a Research Scientist specializing in Generative AI for Physical AI at our Cosmos team, you will develop next-generation algorithms that bridge virtual and physical realms, focusing on advanced video generative models and video-language models. You’ll architect data processing pipelines to generate high-quality training data, design physics simulation algorithms, and optimize large-scale systems using 20,000+ GPUs for foundation model training. Additionally, you will publish influential research papers, collaborate with internal teams and external researchers, and contribute to open-source initiatives. The ideal candidate holds a PhD in Computer Science or related fields, with expertise in PyTorch, diffusion models, vision language, reinforcement learning, physics simulation, 3D generative models, and large-scale distributed training systems. You should also have experience in robotics, successful research-to-product transitions, and contributions to major AI conferences and open-source projects. Join us as we redefine the future of Physical AI with cutting-edge technology and global impact.
Skills
PyTorch
Generative_AI
Physics_Simulation
Reinforcement_Learning
Diffusion_Models
Vision-Language_Models
3D_Generative_Models
Large-Scale_Distributed_Training
CUDA
GPU_Optimization
Open_Source_Contributions
CI/CD
Linux
What you'll do
Pioneer revolutionary generative AI algorithms for advanced video applications.
Architect data processing pipelines to generate high-quality training datasets.
Develop physics simulation algorithms to enhance Physical AI training systems.
Scale large-scale training systems to efficiently use 20,000+ GPUs.
Author influential research papers on groundbreaking discoveries in AI.
Facilitate technology transfer and contribute to open-source initiatives.
Collaborate with diverse teams to drive innovation in Physical AI.
What we're looking for
PhD in Computer Science, Engineering, or related field.
Expertise in PyTorch and libraries for Generative AI and Physical AI.
Strong foundation in diffusion models, vision language, and reinforcement learning.
Knowledge of physics simulation integration with AI systems.
Experience with large-scale distributed training systems and 3D generative models.
Market check
Salary context
This
$168,000–$264,500
range sits above
56%
of similar postings on FindRole.
Peer median band
$168,000–$243,500
Median floor and ceiling across peers.
Typical midpoint (25–75%)
$162,000–$246,150
Middle half of comparable postings.
Based on 239 comparable postings.*
* 240 is the maximum number of comparable postings sampled.
Employer
About Nvidia
Nvidia is a leading designer of graphics processing units (GPUs) and system-on-chip units, powering gaming, professional visualization, data centers, and artificial intelligence workloads. Industry: Semiconductors & AI Computing
Nvidia currently has
802 open roles
on FindRole.
Listed pay typically runs
$184,000–$287,500
across 798 roles with salary data.
Us, Ca, Santa Clara, US139 days ago
$168,000–$264,500
Python
C++
CUDA
PyTorch
3D graphics development
AI algorithms
Computer vision algorithms
Deep learning frameworks
NVIDIA Isaac
Omniverse
Metropolis
NuRec
XR
AI4Media