SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Niantic Spatial is building spatial AI models and services to help people and machines reason, learn, and act safely in the physical world. The company transforms camera imagery, video, sensor data, and archives into accurate, AI-readable representations of real places that robots can train in and machines can use for precise localization.
You'll join the Research Team in London as a 3D Reconstruction Research Scientist/Engineer, working across the full pipeline from cutting-edge academic research to direct customer deployment. The team builds reconstruction technology that converts casually captured or domain-specific imagery into high-fidelity 3D reconstructions, spanning Structure-from-Motion, 3D Gaussian Splatting, feed-forward models, and generative models. Recent work includes shipping 360 splatting to consumers and enterprise customers, developing phone-based splatting technology, and publishing at top venues (CVPR, ICCV, ECCV, SIGGRAPH).
Key responsibilities include conducting original research in 3D computer vision, machine learning, and spatial AI; designing and implementing novel algorithms for 3D reconstruction, neural scene representations, 3D Gaussian Splatting, SfM, meshing, geometric deep learning, and spatial foundation models; transitioning research into customer-facing products under real capture conditions and hardware constraints; building large-scale experiments and benchmarking frameworks; collaborating across computer vision, ML, graphics, and robotics disciplines; and rapidly prototyping state-of-the-art methods for both academic publishing and production deployment.
This role is calibrated to both junior and senior researchers, with scope and responsibility adjusted to experience level. You'll work with world-class researchers on large-scale spatial understanding, localization, reconstruction, scene understanding, and geospatial foundation models.
REQUIREMENTS:
- PhD in Computer Vision, Robotics, Machine Learning, Computer Science, Mathematics, Physics, or closely related field — or equivalent demonstrated research experience in industry
- Strong research track record in 3D computer vision: reconstruction, depth estimation, novel view synthesis, SfM, SLAM, or closely related areas
- Meaningful publication record at top-tier venues (CVPR, ICCV, ECCV, SIGGRAPH, NeurIPS, or equivalent)
- Strong programming skills in Python and hands-on experience with PyTorch
- Evidence of carrying ideas from research into production: shipped features, production pipelines, widely used open-source releases, or similar
- Independence to drive research projects from idea generation through experimentation, publication, and handoff
NICE TO HAVE:
- Experience with classical reconstruction tooling (COLMAP or similar)
- C++ or CUDA experience and understanding of performance and on-device constraints
- Experience with traditional 3D techniques: meshing, texturing, or bringing modern geometry or generative representations into native 3D
