Senior Solutions Architect, Robotics Foundation Model Training

We are building a team of innovators to help partners develop and adopt the next generation of Physical AI, spanning data generation, large-scale multimodal model training, robotics simulation and deployment!

We are looking for a hands-on Applied Engineer with deep expertise in training foundation models at scale and a strong background in robotics.

This role operates at the intersection of innovative AI research, accelerated computing and real-world applications, offering a unique opportunity to work directly with model builders to scale cutting edge Robotics Models from experimentation to production.

Collaboration spans research, engineering, and customer teams, influencing both product direction and applied AI adoption.

Come join us and help shape the future of robotics foundation model training!

Engage with Researchers and ML engineers to architect and optimize end-to-end training workflows for robotics foundation models, like World Models, VLAs, WAMs.

Build proof-of-concepts, reference architectures, and agentic workflows that accelerate experimentation, benchmarking, and model improvement of NVIDIA’s Robotics Open model platforms like Cosmos and GR00T.

Scale pre-training, fine-tuning, and reinforcement learning workloads across multi-GPU and multi-node systems, improving utilization, throughput, and memory efficiency.

Identify and eliminate data pipeline bottlenecks across storage, networking, preprocessing, and data loading for multimodal datasets (video, sensor data, trajectories).

Collaborate with NVIDIA product and engineering teams to provide feedback that shapes future Physical AI platforms

MS, PhD, or equivalent experience in Computer Science, Artificial Intelligence, Electrical or Computer Engineering, Robotics, or a related field.

5+ years of industry or research experience in deep learning, distributed computing, or large-scale model training.

Hands-on experience training or optimizing multimodal or foundation models (e.g., VLMs, VLAs, World Models), ideally in robotics settings.

across the AI model lifecycle, including pre-training, supervised fine-tuning, RL or other post-training methods, evaluation, and model optimization.

Strong expertise in distributed training techniques (data/model/pipeline parallelism, sharding, check-pointing) on multi-GPU or multi-node systems.

Expertise with multimodal training frameworks such as PyTorch, NVIDIA NeMo, JAX, or Hugging Face Transformers.

building or working with high-throughput data pipelines for large-scale training, including storage bandwidth, network throughput, and preprocessing (e.g., decoding, tokenization, batching)

Strong communication skills with the ability to effectively collaborate across Researchers, Engineers and executives.

Ways to Stand Out From the Crowd:

Familiarity with NVIDIA AI and robotics platforms (e.g., Cosmos, GR00T, NeMo, Isaac Sim, Isaac Lab)

with robotics AI workloads, including reinforcement learning in simulation and synthetic data generation.

profiling and optimizing workloads using tools such as Nsight Systems, Nsight Compute, or PyTorch Profiler

Demonstrated impact improving training efficiency and scaling performance

Experience building agentic workflows for automated experimentation, model evaluation, data analysis, or research acceleration

You will also be eligible for equity and benefits.

This posting is for an existing vacancy.

NVIDIA uses AI tools in its recruiting processes.