Frontiers Market - Livestock Mobile App

Ranchers need to manage hundreds of head of cattle across thousands of acres with no reliable cellular coverage. Traditional livestock management requires manual record-keeping, leading to errors, lost data, and inability to track individual animal health and lineage. Existing mobile solutions required constant internet connectivity and failed in rural environments.

Built a native Swift application leveraging iOS Core ML and Vision frameworks for on-device computer vision. The app can identify individual cattle from photos using facial recognition techniques adapted for livestock, storing all data locally in SQLite. Implemented a sophisticated store-and-forward sync mechanism that queues operations and intelligently syncs when connectivity is detected. Ranchers can photograph cattle, record health data, track lineage, and manage inventory entirely offline.

Built and shipped to the App Store, with a few hundred users (~300) and no revenue to date. It handles cattle record-keeping — health, lineage, inventory — backed by a local database with store-and-forward sync that queues offline and syncs when connectivity returns, so it works without a live connection. The AI piece is the on-device computer vision for vision-based weight estimation, which reached ~96% accuracy in controlled tests, though that hasn't yet been matched by in-person field numbers. Broader AI across the record-keeping is planned, not built.