SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
SpAItial is pioneering next-generation World Models, advancing generative AI, computer vision, and physics-aware 3D simulation. The company is moving beyond 2D pixels to build models that natively understand the physics and geometry of the real world, with applications across robotics, AR/VR, gaming, and cinema.
As a Research Engineer focused on 3D Reconstruction, you will drive improvements in reconstruction quality from research conception through production implementation. You'll work across the full spectrum of reconstruction challenges: geometry and appearance modeling, material and lighting estimation, handling sparse or incomplete observations, dynamic scene capture, and large-scale environment reconstruction.
Key responsibilities include:
- Advancing 3D reconstruction quality by translating research ideas into working methods
- Implementing and improving reconstruction approaches across the full range, from per-scene optimization to feed-forward neural models
- Staying current with frontier scene representations (Gaussian splats, radiance fields, etc.), implementing new techniques as they emerge, and extending or adopting them where they provide advantages
- Tuning losses, priors, and optimization strategies to improve fidelity, robustness, and computational efficiency
- Building rigorous evaluations using public and internal benchmarks to drive technical decisions
- Collaborating with research and engineering teams to integrate successful methods into production systems
This is a hands-on role requiring deep reconstruction expertise and strong software engineering fundamentals.
QUALIFICATIONS:
- Bachelor's or Master's degree (or equivalent experience) in computer science, computer vision, graphics, or related field
- Deep, hands-on knowledge of modern scene representations (e.g., Gaussian splats, radiance fields) and ability to evaluate their trade-offs
- Hands-on experience across the reconstruction spectrum, from per-scene optimization to feed-forward models
- Solid understanding of multi-view geometry, camera models, and rendering
- Strong experience with deep learning frameworks, including designing and defending large-scale experiments and comparisons
- Strong coding skills and demonstrated ability to implement methods from research papers into working systems
- PhD or publications in reconstruction are a plus, not required
