This role pays more than
59%
of similar roles. Most pay
$213,375–$256,575
— the shaded band above.
At the midpoint, this role pays about
$248k
versus about
$235k
for comparable roles.
Based on 240 similar postings.
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
896 open roles
on FindRole.
Listed pay typically runs
$184,000–$287,500
across 876 roles with salary data.
As a Senior Research Scientist, Multimodal Foundation Models and Robotics within the Generalist Embodied Agent Research group, you will develop foundation models and systems for humanoid robots and embodied agents. Your daily responsibilities include designing and implementing novel AI algorithms, developing large-scale training and inference methods for foundation models, and optimizing model deployment across physical simulation environments and robot hardware. You will work with technologies including PyTorch, Jax, TensorFlow, Python, C++, CUDA, ROS, and physics simulation frameworks like MuJoCo and Isaac Sim. The role focuses on the technical challenge of creating general-purpose agents capable of mastering complex skills in both virtual and physical worlds. Key research areas include large vision-language models, video generative models, diffusion algorithms, action-based transformers, reinforcement learning, imitation learning, robot kinematics, dynamics, and various control methods like PID and model predictive control.
What does a Research Scientist earn?
Median $216250 from 42 postings across 13 companies.
Design and implement novel AI algorithms for humanoid robots and embodied agents.
Develop large-scale training and inference methods for multimodal foundation models.
Optimize and deploy AI models within physical simulations and on robot hardware.
Conduct research in areas like LLMs, vision-language models, and action-based transformers.
Implement robot learning techniques including reinforcement learning and imitation learning.
Manage the integration of robot kinematics, dynamics, and sensor data into systems.
Operate and test robot hardware and lab equipment safely.
Utilize physics simulation frameworks like MuJoCo and Isaac Sim for model development.
What we're looking for
A Ph.D. in Computer Science, Engineering, Electrical Engineering, or equivalent research experience is required.
Candidates must have 5 years of relevant work or research experience in Multimodal Foundation Models and/or Robotics.
Experience with LLMs, vision-language models, video generative models, diffusion algorithms, or action-based transformers is required.
Proficiency in Python is required, with C++ and CUDA skills being a significant plus.
Candidates must have hands-on experience and publications in robot learning, including reinforcement learning, imitation learning, or classical control.
Expertise in large-scale machine learning systems, compute infrastructure, and training frameworks like PyTorch, Jax, or TensorFlow is required.
Deep understanding of robot kinematics, dynamics, sensors, and control methods such as PID and model predictive control is required.
Familiarity with physics simulation frameworks like MuJoCo and Isaac Sim, along with experience in robot hardware design, is required.
Python
C++
CUDA
PyTorch
JAX
ROS2
NVIDIA Isaac Sim
Isaac Lab
MuJoCo
Warp
Reinforcement Learning
Imitation Learning
Task and Motion Planning
Sim-to-Real
Computer Vision
VLMs
CI/CD
Version Control
Python
Rust
C++
ROS
Mujoco
Isaac
Gazebo
MLOps
Physics Simulation
Sensor Fusion
Teleoperation
Control Algorithms
Robotics
Hardware Deployment
Data Processing
Python
C++
CUDA
PyTorch
TensorFlow
Machine Learning
Robotics
Autonomous Vehicles
Parallel Programming
VLMs
Closed-loop Training
AI Safety
Simulation
open-source
PyTorch
NVIDIA NeMo
JAX
Hugging Face Transformers
Distributed Training
Reinforcement Learning
Multi-GPU
Multi-node
Isaac Sim
Isaac Lab
Nsight Systems
Nsight Compute
PyTorch Profiler
Data Pipelines
Model Optimization
Synthetic Data Generation
VLMs
VLAs