Software Engineer, Robotics Foundation Models (Contractor) – NVIDIA Robotics Lab

About this role

Software Engineer, Robotics Foundation Models (Contractor) – NVIDIA Robotics Lab
Role Summary

The
NVIDIA Robotics Lab
is seeking high-caliber
Contractor Software Engineers
to join our Robotics Foundation Models team onsite in Santa Clara, CA. This is a high-impact,
contract-based role
focused on bridging frontier research with production-grade engineering..

We are looking for versatile engineers to specialize in one or more of these four core pillars:

1. Fullstack & Visualization:
Building the interfaces that make robotics data actionable.
2. Model Optimization:
Improving training and inference efficiency across massive GPU clusters.
3. AI Infrastructure & MLOps:
Scaling the pipelines that power our research velocity.
4. Data Quality & Automation:
Engineering the "Data Factory" for Embodied AI.

Key Responsibilities by Focus Area
1. Fullstack & Operational Intelligence

• Architect and scale high-performance dashboards to visualize high-dimensional robotics metrics, fleet telemetry, and research experiment data.
• Design and develop features in Python and TypeScript to stream complex robotics models outputs to web applications.
• Build resilient APIs and manage complex database migrations to ensure data integrity for massive datasets.

2. Model Optimization

• Partner with researchers to profile and optimize model training and inference performance across distributed GPU clusters.
• Implement strategies for cross-modality transfer, image encoding, and tokenization efficiency.
• Develop low-level performance profiling to maximize GPU utilization and training throughput.

3. AI Infrastructure & MLOps

• Design and maintain high-throughput developer tools and automated frameworks specifically for training and evaluating robotics foundation models.
• Create workflows (Spark, Ray, GCP) to download and preprocess petabyte-scale image, text, and robotics sensor data.
• Develop comprehensive test frameworks ...