Staff Graphics Engineer - 4D Reconstruction & GPU Rendering

Avala; San Francisco (CA); Hybrid; USD 180,000 - 250,000; Full time

Job summary

Avala in San Francisco is seeking a senior architect to lead the 4D reconstruction pipeline, fusing LiDAR, camera, radar, and IMU data into physics-grade ground truth. You will own the Gaussian Splatting renderer, temporal tracking, and scalable compute infrastructure.

You'll guide the rendering team, mentor engineers, and collaborate with ML engineers on auto-labeling and scene understanding models to push autonomous systems forward.

Qualifications

• 7+ years in graphics, 3D vision, or high-performance computing.
• Deep expertise in GPU programming (WebGPU, CUDA, Vulkan, or Metal).
• Strong C++, Rust, or Dart performance optimization experience.
• Published work or production systems in 3D reconstruction or novel view synthesis.
• Experience leading technical teams of 3+ engineers.

Responsibilities

• Lead development of the 4D reconstruction pipeline (Gaussian Splatting, NeRF, multi-view stereo).
• Design GPU-accelerated compute shaders for real-time point cloud processing and scene rendering.
• Build temporal tracking and object persistence systems across sensor frames.
• Architect distributed processing pipelines that scale to fleet-level data volumes.
• Set technical direction for the rendering team and mentor senior engineers.
• Collaborate with ML engineers on auto-labeling and scene understanding models.

Skills

Graphics 3D experience

GPU programming

C++ optimization

Rust optimization

Dart optimization

Leadership

Tools

WebGPU

CUDA

Vulkan

Metal

Job description

Avala in San Francisco is seeking a senior architect to lead the 4D reconstruction pipeline, fusing LiDAR, camera, radar, and IMU data into physics-grade ground truth. You will own the Gaussian Splatting renderer, temporal tracking, and scalable compute infrastructure.

You'll guide the rendering team, mentor engineers, and collaborate with ML engineers on auto-labeling and scene understanding models to push autonomous systems forward.