I want to turn large sets of aerial images into accurate, map-ready datasets through an AI-driven photogrammetry workflow. The core of the job is data analysis: training and integrating models that can detect tie points, correct distortions, and automatically build orthomosaics and dense point clouds. The end product will serve cartography and geomatics needs, so georeferencing accuracy and the ability to export into common GIS formats (GeoTIFF, LAS/LAZ, shapefile) are essential. I will supply representative image batches and the ground-control information; you guide the tooling. If you prefer Python with OpenCV, PyTorch or TensorFlow, great—just keep the pipeline reproducible and well documented. A concise README plus commented code that I can spin up on a standard cloud VM will be the expected hand-off.