SynthBook — ML-Powered Book Recommendation System
Python
Django
scikit-surprise
scikit-learn
Pandas
NumPy
TF-IDF
Cosine Similarity
FunkSVD
SQLite
WhiteNoise
Gunicorn
About this project
SynthBook is a full-stack Django book recommendation engine trained on the 6M-rating goodbooks-10k dataset. It uses a hybrid algorithm combining FunkSVD for personalized embeddings, TF-IDF for metadata matching, and the IMDB weighted formula for global and genre rankings. By loading pre-trained models at startup, the backend delivers instant, real-time recommendations across four tiers with zero per-request retraining. The system features django-allauth authentication and is deployed on Render via automated GitHub CI/CD.