Reels is a full-stack media platform designed to handle complex user data life-cycles, secure session management, and persistent viewing states. This project served as the architectural foundation and inspiration for the collaborative project K-FEELZ.

• Relational Data Modeling: Architected a scalable PostgreSQL schema using Prisma to manage high-volume user interactions, including viewing history and watchlists with unique constraints to ensure data integrity.
• Security & Session Rigor: Engineered a professional-grade security layer using JWT and Bcrypt. Developed custom Middleware to extract and validate tokens from secure cookies, implementing graceful error handling for unauthorized access.
• Asynchronous State Management: Utilized Redux Toolkit and createAsyncThunk to manage non-deterministic API lifecycles (pending, fulfilled, rejected), ensuring the UI remains responsive during high-latency media fetching.
• Media Preview Logic: Developed a dynamic preview system that utilizes conditional rendering to provide users with a video/modal preview of media assets before engagement.

• Frontend: React, Redux Toolkit, CSS3 (Flexbox/Grid)
• Backend: Node.js, Express.js
• Database: PostgreSQL, Prisma ORM
• Auth: JSON Web Tokens (JWT), Bcrypt, Cookie-parser
• APIs: The Movie Database (TMDB) API

The application utilizes a custom authenticateToken middleware that acts as a gatekeeper for all media routes. By verifying signatures against a protected JWT_SECRET, the system prevents unauthorized data tampering.

Leveraging Prisma's upsert logic, the viewing history tracker ensures that user engagement is recorded in real-time. If a user re-watches a title, the system updates the watchedAt timestamp rather than creating redundant entries, maintaining a clean and performant database.

Reels provided the core authentication and media fetching logic that was later scaled into K-FEELZ. The transition involved moving from a search-based discovery model to a weighted, logic-driven recommendation engine.