Staff Software Engineer - Rendering (Simulation)

Job Description

The Rendering team builds and maintains the core sensor simulation system that produces physically accurate synthetic sensor data for autonomous vehicle development. We own the full rendering pipeline — from scene ingestion and acceleration structure construction through GPU ray tracing and sensor-specific post-processing. Our stack includes C++, CUDA, NVIDIA OptiX, USD (Universal Scene Description), MDL materials, and ROS. We care deeply about performance, correctness, and clean architecture.

The Role

As a Staff Software Engineer on the Rendering team within Simulation, you will architect, optimize, and extend a GPU-accelerated, physics-based sensor simulation system used to generate synthetic sensor data — cameras, LiDAR, radar, and depth sensors — for autonomous vehicle development and validation. You will work at the intersection of real-time rendering, GPU computing, and large-scale scene management, tackling challenges in memory management, geometry streaming, material systems, and ray tracing performance. Your work will directly impact the fidelity, scalability, and speed of the simulation platform that AV teams depend on for training, testing, and validating perception and planning systems.

What You'll Do

- Design and implement systems for efficiently loading, managing, and rendering large-scale 3D scenes — including geometry streaming, level-of-detail, instancing, and acceleration structure optimization.

- Optimize GPU memory management and texture pipelines to support scenes with significantly higher geometric and material complexity.

- Architect and extend the sensor simulation pipeline to support new sensor modalities and integrate them into our ROS-based data delivery stack.

- Improve ray tracing performance through BVH optimization, shader efficiency, and pipeline profiling using CUDA and OptiX.

- Enable new use cases such as real-time or near-real-time rendering for reinforcement learning workloads, requiring low-latency scene updates and rapid frame generation.

- Contribute to the USD scene ingestion and material translation pipeline, improving robustness, performance, and support for complex production assets.

- Drive architectural improvements across the codebase — refactoring, API design, build system improvements, and developer tooling.

- Uphold high code quality and architecture standards through code reviews, technical design documents, and mentorship of other engineers.

- Leverage AI-assisted development tools to accelerate prototyping, testing, and code quality.

Your Skills & Abilities

- BS, MS, or PhD in Computer Science, Computer Graphics, Engineering, or equivalent experience.

- 7+ years of professional C++ development experience, with deep knowledge of modern C++ (C++17/20), memory management, and performance optimization.

- Significant experience with GPU programming (CUDA, OptiX, Vulkan, or similar) and an understanding of GPU architecture and execution models.

- Proven track record working on rendering systems, game engines, simulation platforms, or other performance-critical real-time graphics software.

- Strong understanding of data structures and algorithms relevant to rendering: spatial acceleration structures (BVH/kd-tree), scene graphs, texture management, and resource streaming.

- Experience profiling and optimizing complex C++ and GPU codebases (e.g., Nsight, Tracy, perf, custom instrumentation).

- Comfort working in large, multi-module codebases with CMake-based build systems.

- Strong communication and collaboration skills, with demonstrated ability to drive technical design across teams.

What Will Give You A Competitive Edge

- Experience with USD (Universal Scene Description) or similar scene description frameworks.

- Experience with physically-based rendering, material systems (MDL, OSL), or spectral rendering.

- Familiarity with ROS and integrating rendering or sensor simulation into robotics data pipelines.

- Experience with reinforcement learning environments or gym-style simulation interfaces.

- Knowledge of ray tracing theory, radiometric sensor modeling, or computational