We need a cross-domain and a single-domain recommendation engines (includes algorithm implementation and REST APIs) that can make recommendations from one item space to the same item space or another item space. We need to implement it in Book and Movies domain. For example, if user likes a book, we should be able recommend similar books as well as similar movies too, and vice versa. We do not need professional testing for this project, even a POC quality will be acceptable. We do not require an UI at this point. Feel free to use any open data available on books and movies.