Design a Book Recommendation System

Design a book recommendation system. Users and publishers can create posts (reviews, recommendations) that have likes and comments; users can follow each other. Users receive recommendations both from publishers they follow and from the system. Cover the data model for posts/likes/comments and their aggregation, and — the key focus — the system-generated recommendation engine: scoring books across genres and modeling each user''s genre interests (e.g. by mining posts), then a recommender service that generates posts for a user''s timeline. Discuss aggregation at scale.