Research Scientist 3D Reconstruction SfM and SLAM
SpAItialVisit SpAItial website
SpAItial is an AI company building physically grounded world models that generate persistent, explorable 3D worlds from text, images, and panoramas.
London, United Kingdom
Funding history
About SpAItial
SpAItial Ltd operates the Echo model family, a spatial-AI product available through a web app and developer API for generating, editing, sharing, and exporting 3D Gaussian Splat worlds.
Skills
About the Role
You will develop robust methods to recover camera poses and geometry from real-world imagery. You will build and improve structure-from-motion, camera-tracking, bundle-adjustment, and learned reconstruction systems, evaluate their quality, scale them to diverse datasets, and help bring advances into production.
Requirements
- PhD in computer vision focused on 3D reconstruction
- Publications at top venues including CVPR, ICCV, ECCV, or NeurIPS
- Deep understanding of multi-view geometry, camera models, epipolar geometry, triangulation, and PnP
- Familiarity with VGGT, DA3, Pi3, deep MVS, learned matching, and monocular depth estimation
- Hands-on experience with SfM or SLAM systems including COLMAP or ORB-SLAM and nonlinear least-squares solvers including Ceres
- Strong Python and PyTorch skills
- Ability to debug failure cases on challenging real-world data
Responsibilities
- Design camera-pose estimators and 3D reconstructors
- Build robust SfM and camera-tracking pipelines for varied imaging sensors
- Develop bundle adjusters and nonlinear optimizers, including non-perspective camera formulations
- Integrate and extend feed-forward reconstructors including VGGT, DA3, and Pi3
- Advance deep multi-view stereo, learned matching, and monocular-depth methods
- Build pose-accuracy and reconstruction-quality metrics and improve benchmark performance
- Scale reliable and efficient reconstruction methods to large real-world datasets
- Collaborate to bring reconstruction advances into production systems