Photo memory app with face detection and automatic album generation as core differentiators. Built async processing pipeline keeping upload experience fast while ML models run in background. Row-level security in PostgreSQL enforces privacy as structural guarantee.
🎯 Overview
Momentz organizes personal photo memories by recognized faces, automatically grouping images across uploads. Async processing returns immediately to users while face detection and collection generation complete independently. Privacy-first architecture prevents cross-user data leaks.
🛠️ Tech Stack
- C#: Async/await upload handler for concurrent photo uploads with fire-and-forget job dispatch.
- Python: Face detection and facial embedding extraction via OpenCV running as independent worker consuming job queue.
- OpenCV: Proven face detection accuracy on varied lighting and angle conditions typical of personal photos.
- AWS S3: Durable media storage for original and compressed assets decoupled from API servers.
- PostgreSQL: Row-level security enforcing per-user data isolation at database level for privacy.
📈 Key Features
- Fast uploads with background face detection and album generation
- Automatic face recognition grouping photos across uploads
- Secure photo storage with media compression optimization
- Row-level security enforcing privacy structurally at database level
- Timeline organization with metadata tagging and sharing controls