SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Mecka AI is building the data infrastructure layer for robotics and embodied AI, designing and operating global systems for data capture, labeling, and hardware-enabled workflows used by leading AI labs and robotics companies to train and validate humanoid and embodied AI systems.
You will own the architecture for generating 3D objects from video. Given egocentric footage of someone manipulating an object, you will build models that recover its geometry and ultimately produce a simulatable asset with physical properties suitable for physics simulation. You'll tackle the hardest, most open part of the objects effort: reconstructing objects that are small in frame, partly occluded by the hand, and sometimes deformable.
Key responsibilities include:
- Neural 3D reconstruction: develop models that recover object geometry from monocular/egocentric video using sequential and multi-view approaches robust to occlusion and low resolution
- Generative refinement: use generative 3D to complete and clean up partial reconstructions while rejecting hallucinated geometry
- Toward simulatable objects: move beyond geometry toward assets that carry physical properties and rigging so reconstructed objects can be simulated
- Contact-aware reconstruction: use hand contact as a signal to constrain object shape and pose
- Deformables: address the frontier challenge of objects that change shape as they're handled
- Prototype and hand off: build research prototypes and transfer them to the integrations researcher for productization; publish where appropriate
You'll own an open, unsolved problem on real, large-scale data—not a fully-specified spec. You'll push the frontier, then hand results to the integrations researcher who takes them into the production pipeline. Success means manipulated objects reconstruct from egocentric video at usable quality (including small, occluded, and deformable cases), a credible path exists from interaction video to simulatable 3D objects, and your research transfers into the pipeline through the integrations seat.
REQUIREMENTS:
- Strong 3D/neural-reconstruction research background: neural implicit or explicit 3D, 4D reconstruction, or generative 3D experience at the frontier
- Deep learning depth: design and train models; read and reproduce frontier papers
- 3D geometry foundation: structure-from-motion, multi-view geometry, meshes, and differentiable rendering
- Research taste: ability to find tractable decomposition of hard, open problems and make steady progress
- Engineering capability: prototype fast in clean Python/PyTorch; get research ideas working on real data
Strong plus:
- Generative 3D (diffusion/feed-forward 3D-generation families)
- Differentiable rendering, NeRF, or Gaussian splatting
- 4D/dynamic-scene reconstruction from video
- Physics simulation and rigging
- Hand-object interaction; publications at top vision or graphics venues
Tech stack: Python/PyTorch (primary) for model design and training; 3D deep learning (neural implicit/explicit representations, differentiable rendering); generative-3D toolkits; mesh processing (Open3D, trimesh); physics simulators.
