PhD - Neural 3D Scene Representations for the Automotive Cabin
Core
Researching neural 3D scene representations (NeRF, Gaussian splatting, implicit occupancy) to reconstruct vehicle occupants from fixed camera viewpoints for safety and comfort systems.
Role type
PhD researcher in 3D computer vision and generative modeling
Builds
Unified 3D representations of vehicle interior and exterior for end-to-end ADAS systems
Domain
Automotive AI, 3D computer vision, generative modeling
Deliverable
production ML models
Required skills
machine learning with PyTorch, vision-based models, neural scene representation, Python programming, computer vision, 3D geometry, camera models, projective geometry, multi-view constraints
Preferred skills
3D scene representations (NeRF, Gaussian splatting, implicit/occupancy networks), depth/3D estimation, differentiable rendering, articulated human pose estimation
Technologies
PyTorch, NeRF, Gaussian splatting, diffusion models
Responsibilities
Investigate neural 3D scene representations adapted to in-cabin perception challenges; leverage known cabin geometry as built-in supervision; incorporate 3D-aware diffusion models as learned priors; reconstruct occupants from one to three fixed camera viewpoints; collaborate with academia and industry experts.
Seniority
PhD candidate, research-focused
