AI Movie Recommendation System

Production-grade recommender system for a streaming-style use case, designed around sparse feedback, cold start behavior, leakage-aware evaluation, and serving reliability.

Highlights

• Built a hybrid recommendation pipeline using explicit and implicit feedback signals.
• Used a temporal train/validation/test split to evaluate the model without leakage.
• Served recommendations through a Flask API with CI/CD, Docker, testing, and telemetry.
• Added graceful degradation with an ALS fallback model for reliability.

Tech Stack

Python, scikit-learn, Flask, Kafka, pytest, Docker, and GitHub Actions.

Why It Matters

This project demonstrates end-to-end ML system design, not just modeling, with attention to evaluation rigor, service reliability, and production-readiness.