3D Vision Engineer: Real-Time SLAM, NeRFs & Spatial AI

Avala; San Francisco (CA); Hybrid; USD 130,000 - 190,000; Full time

Benefits offered by this job

Equity

Health insurance

Vision care

Dental coverage

FSA

Unlimited PTO

Job summary

Avala in San Francisco is seeking a 3D Vision Engineer to develop 3D reconstruction and scene understanding for autonomous vehicles and robotics. You will build robust perception systems using NeRFs, Gaussian Splatting, SLAM, and scale our 4D labeling platform.

The role emphasizes production-scale processing, collaboration with the annotation team, and real-time SLAM, multi-view geometry, and deep learning with PyTorch/TensorFlow.

Qualifications

• 3+ years in 3D vision, graphics, or spatial computing.
• Strong linear algebra and computational geometry fundamentals.
• Deep learning expertise with PyTorch or TensorFlow.
• Multi-view geometry and rendering pipeline experience.
• Published research or shipped production systems in 3D reconstruction.

Responsibilities

• Develop 3D reconstruction pipelines using NeRFs, Gaussian Splatting, and multi-view stereo.
• Build real-time SLAM systems for autonomous vehicle and robotics perception.
• Create scene understanding models that parse complex driving and industrial environments.
• Implement spatial computing techniques for our 4D annotation platform.
• Optimize 3D pipelines for production-scale processing of fleet sensor data.
• Collaborate with the annotation team to improve 3D auto-labeling accuracy.

Skills

3D vision

Linear algebra

Deep learning (PyTorch/TensorFlow)

Multi-view geometry

SLAM & 4D sensing

Tools

ROS

LiDAR processing

Real-time rendering

Game engines

Job description

Avala in San Francisco is seeking a 3D Vision Engineer to develop 3D reconstruction and scene understanding for autonomous vehicles and robotics. You will build robust perception systems using NeRFs, Gaussian Splatting, SLAM, and scale our 4D labeling platform.

The role emphasizes production-scale processing, collaboration with the annotation team, and real-time SLAM, multi-view geometry, and deep learning with PyTorch/TensorFlow.