Building an AI Book Discovery Experience with FastAPI and Supabase
A behind-the-scenes look at recommendation logic, UI polish, and the tradeoffs involved in shipping a full-stack reading product.
Why This Project Was Interesting
Book recommendation products sit at the intersection of data, interface design, and user trust. A recommendation engine can be technically correct and still feel unhelpful if the browsing experience is dull or confusing.
I wanted this project to feel useful and personal, not like a search result page with a few filters attached.
System Design Choices
The frontend focused on smooth browsing, clear hierarchy, and motion that made discovery feel lightweight. On the backend, FastAPI handled recommendation logic while Supabase supported authentication and data persistence.
That split let me keep the product responsive while still experimenting with recommendation quality and personalized features.
What I Learned
A polished product comes from the connection between systems, not from any single feature. Recommendation quality, loading states, onboarding, and saved content all shape whether a user feels the app understands them.
This project helped me think more carefully about how machine learning features should be presented in a human-centered product.