Computer Vision Research Engineer at stealth AI-native design startup

You will advance indoor spatial reconstruction using smartphone images, video, and depth data. This role involves researching and implementing reconstruction and scene-understanding methods for both LiDAR and non-LiDAR capture. You will bridge the gap between research and production, building core models for 3D scene reconstruction, material recognition, and photorealistic synthesis for physical spaces.

Stealth AI-native design startup

• Develop and implement state-of-the-art methods for photo-to-3D scene reconstruction, including NeRF and Gaussian splatting techniques.
• Design and execute experiments for scene segmentation, material recognition, and accurate measurement from ordinary smartphone imagery.
• Build benchmarks and integrate high-performance vision models into production, owning the entire pipeline from data to deployment.

• Proven research depth in 3D computer vision and multi-view geometry, with experience designing and training novel vision models.
• Strong proficiency in PyTorch and the engineering discipline to move research into production environments rather than just notebooks.
• High comfort level with early-stage startup ambiguity and a desire to take end-to-end ownership of complex problem areas.

• Own a greenfield CV/3D research agenda at the core of a product defining the future of physical space design.
• Work within a high-caliber team of approximately 18 engineers at a stealth startup backed by top-tier venture capital.
• Direct line to the CEO with the opportunity to see your research models shipped directly to enterprise and boutique customers.