Senior Synthetic Data Engineer - Autonomous Driving

NVIDIA is building the next generation of simulation and synthetic data workflows through NVIDIA DRIVE, with a focus on producing high-quality multi-sensor datasets for autonomous driving training. In this role, you will design and develop simulation environments and synthetic data pipelines using NVIDIA Omniverse NuRec and Cosmos, supporting both sensor simulation and dataset evaluation at scale. The work spans reconstruction-driven worlds, controllable scenario generation, and the engineering needed to run production-grade pipelines in data center or cloud environments.

What you’ll do

• Build, implement, and optimize tools that generate synthetic data for training different DRIVE deep learning networks, covering simulated lidar, radar, camera/RGB-D, bounding boxes, object tracks, world models, segmentation, depth, scene semantics, and sensor metadata.
• Develop lidar and radar sensor simulation workflows that operate on NuRec reconstructed driving worlds and Cosmos-generated environments, including sensor placement, calibration, material response, geometry handling, noise modeling, and scenario variation.
• Develop a Cosmos world model to improve world generation, including controllable scenario generation, novel view synthesis, trajectory extrapolation, scene completion, quality triage, regression detection, and controllability evaluation.
• Gather perception, planning, and DRIVE network requirements, then align them with current synthetic data and sensor simulation capabilities, creating new tools or improving existing performance when gaps appear.
• Create dataset quality assessments and synthetic-real comparison procedures to evaluate sensor realism, annotation quality, distribution coverage, scenario diversity, and sim-to-real transfer for autonomous driving.
• Set up, profile, and oversee large-scale NuRec, Cosmos, and sensor simulation pipelines in data center or cloud environments.
• Debug cross-stack systems across sensors, reconstruction models, world models, simulation runtime, GPU workloads, distributed data services, and downstream autonomous-driving workloads.

Qualifications

• B.S. or M.S. in Computer Science, Electrical Engineering, Computer Engineering, Applied Math, Physics, or a related field (or equivalent experience).
• 8+ years of experience in computer graphics, computer vision, autonomous driving, sensor simulation, neural rendering, or physically-based sensor modeling, synthetic data generation, or closely related software engineering roles.
• Strong Python and C++ skills, with experience building, debugging, profiling, and maintaining production-quality systems on Linux.
• Solid mathematical foundation in linear algebra, geometry, and probability.
• Familiarity with synthetic data annotations, data formats, dataset curation, data augmentation, and evaluation workflows for perception model training and validation.
• Familiarity with deep learning workflows and modern ML tooling, including the ability to translate network needs into synthetic data requirements and measurable quality criteria.
• Experience with scalable engineering workflows such as Git, Docker, Kubernetes, CI/CD, distributed storage, and deployment in data centers or cloud environments.

Technologies

• Python, C++, Linux
• Git, Docker, Kubernetes, CI/CD
• NuRec, Cosmos, NVIDIA Omniverse
• Distributed storage, cloud, data center

Compensation and eligibility

Salary: USD 184,000 - 356,500 per year. You will also be eligible for equity and benefits.

Ways to stand out

• Practical experience working directly with NVIDIA NuRec, Cosmos, world foundation models, Real2Sim systems, or autonomous-driving simulation and validation pipelines.
• Experience in NuRec world reconstruction, neural rendering, 3D Gaussian Splatting, NeRFs, or occupancy networks.
• Deep lidar or radar simulation expertise, including ray tracing or ray casting, reflectance and intensity modeling, Doppler, radar cross-section, weather effects, occlusion, and sensor-specific noise models.
• Experience developing synthetic data pipelines for autonomous driving, including closed-loop simulation, domain randomization, long-tail scenario mining, or sim-to-real transfer.
• Familiarity with autonomous vehicle data pipelines such as OpenDRIVE, HD maps, scenario formats, vehicle dynamics, or AV safety validation.

Location: Santa Clara, CA (onsite).