Computer Vision Engineer

Description:

We aren't just looking for standard object detection. The core of this role is the 3D geometry: taking a 2D pixel from the camera feed and translating it into precise, real-world map coordinates $(X, Y)$ over non-flat terrain.

🛠️ What You Will Build

Detect & Track: A pipeline to detect and maintain stable IDs for many people and vehicles at long distances.

Pixel-to-Coordinate Engine: The geometric software that turns image positions into real-world GPS/map coordinates.

Trajectory Smoothing: Time-based filtering (like Kalman filters) so asset positions stay steady and don't jump around on the map.

Offline Deployment: A self-contained pipeline running on-site that outputs an annotated video feed and a live position log.

🎯 Requirements

5+ years of experience shipping production computer vision software.

Excellent grasp of camera geometry, calibration, and 3D reasoning.

Strong, self-directed software engineering skills and clear communication.

Nice to have: Experience with geospatial/mapping data, state estimation, or long-range outdoor cameras.