Senior Solutions Architect, Robotics Foundation Model Training
Core
Architect and optimize end-to-end training workflows for robotics foundation models (World Models, VLAs, WAMs) and build proof-of-concepts to scale NVIDIA's Robotics Open model platforms.
Role type
Senior Applied Engineer (Robotics Foundation Model Training)
Builds
Robotics foundation models (World Models, VLAs, WAMs) and agentic workflows for NVIDIA's Cosmos and GR00T platforms
Domain
Robotics, Physical AI, Large-scale Multimodal Model Training
Deliverable
production ML models
Required skills
Deep learning, distributed computing, large-scale model training, multimodal/foundation model training (VLMs, VLAs), distributed training techniques (data/model/pipeline parallelism), multimodal training frameworks (PyTorch, NeMo, JAX, Hugging Face), high-throughput data pipelines
Preferred skills
NVIDIA AI/robotics platforms (Cosmos, GR00T, Isaac Sim), reinforcement learning in simulation, synthetic data generation, workload profiling (Nsight Systems/Compute), agentic workflows for automated experimentation
Technologies
PyTorch, NVIDIA NeMo, JAX, Hugging Face Transformers, Nsight Systems, Nsight Compute, Isaac Sim, Isaac Lab, Cosmos, GR00T
Responsibilities
Engage with researchers to architect training workflows for robotics foundation models; Build proof-of-concepts and reference architectures; Scale pre-training, fine-tuning, and RL workloads across multi-GPU systems; Identify and eliminate data pipeline bottlenecks; Collaborate with product teams to shape future Physical AI platforms
Seniority
Senior, hands-on IC
