Computer Vision Engineer (3D / Spatiotemporal AI) – Next-Gen Animal Intelligence Platform

We’re building something that doesn’t exist yet. At NeuralX, we are pushing the frontier of 3D computer vision and spatiotemporal AI to understand animal behavior and biomass in real-world environments — not in controlled labs, but in complex, dynamic, underwater ecosystems. This is not a typical “train a model on a clean dataset” role. You’ll be working on noisy, real-world, multi-camera data, solving problems that sit at the intersection of: 3D reconstruction Multi-view geometry Temporal modeling Behavioral understanding Applied AI in physical environments If you like working on problems where the data is imperfect, the physics matters, and the solution isn’t obvious, you’ll enjoy this. What You’ll Work On Fine-tuning and improving existing models for: 3D biomass estimation 3D behavioral analysis of animals (trajectory, interaction, patterns) Multi-camera calibration and synchronization challenges Spatiotemporal modeling (tracking + sequence understanding) Handling underwater-specific constraints (visibility, distortion, occlusion) Improving robustness in production-like environments Ideal Background You don’t need to check every box, but strong candidates typically have: Solid experience in computer vision / deep learning Hands-on work with: 3D vision (SfM, MVS, NeRF, depth estimation, etc.) Object tracking / multi-object tracking Video understanding / temporal models Strong PyTorch (or equivalent) experience Ability to debug and iterate in messy, real-world datasets Bonus points: Experience with underwater / low-visibility environments Familiarity with geometry-heavy pipelines Experience deploying models in production or near-production settings Why This Is Interesting You’ll work on real-world impact problems (not benchmark chasing) The system combines physics + AI + geometry, not just end-to-end black boxes Opportunity to shape the core intelligence layer of the product Small, highly technical team — fast iteration, high ownership Direct exposure to cutting-edge applications of spatial + temporal AI Engagement Flexible (project-based → long-term possible) Remote-friendly, async collaboration We care more about capability and thinking than credentials To Apply Please include: Relevant projects (GitHub, papers, demos) Brief explanation of a hard vision problem you solved Your experience with 3D or temporal modeling (if any) If you’re excited by solving unsolved problems in the wild, we should talk.