Technical Lead, 3D Vision & Spatial AI

Niantic; San Francisco (CA); Hybrid; USD 180,000 - 240,000; Full time

Job summary

Niantic is seeking a Tech Lead for the Applied Computer Vision Algorithms Team to advance 3D reconstruction, SfM, and Gaussian splatting within our geospatial AI platform. You will translate cutting-edge research into production-grade features and mentor a team of CV engineers in a high-growth environment.

You will collaborate with R&D and product teams, optimize GPU-accelerated code in C++ and Python, and help define the team's technical roadmap while delivering robust, real-world spatial

Qualifications

• 8+ years of professional experience in Computer Vision or Machine Learning (or 6+ years with a PhD).
• Proficient in production-level C++ and Python.
• Strong background in 3D CV/ML, SfM, 3D reconstruction, Gaussian splatting.
• Experience optimizing algorithms for GPUs in Android/Linux environments.
• Previous tech leadership in a CV/ML team in high-growth settings.

Responsibilities

• Translate cutting-edge research papers into production-grade features.
• Lead design and implementation of 3D reconstruction pipelines (SfM, meshes, Gaussian splats).
• Develop and optimize Gaussian Splatting quality algorithms for CPU/GPU.
• Write and maintain high-performance, shader-based production code in C++ for Android and Linux.
• Define technical roadmap and provide mentorship and code governance.

Skills

C/C++

Python

3D Computer Vision

Gaussian Splatting

SfM

GPU optimization

Leadership

Education

Bachelor's degree in CS/Engineering

Master's or PhD preferred

Tools

CUDA

Shader programming

OpenGL

Job description

Niantic is seeking a Tech Lead for the Applied Computer Vision Algorithms Team to advance 3D reconstruction, SfM, and Gaussian splatting within our geospatial AI platform. You will translate cutting-edge research into production-grade features and mentor a team of CV engineers in a high-growth environment.

You will collaborate with R&D and product teams, optimize GPU-accelerated code in C++ and Python, and help define the team's technical roadmap while delivering robust, real-world spatial