Principled and Scalable Diversity-Aware Retrieval via Cardinality-Constrained Binary Quadratic Programming

Build a lightweight Python library that implements CCBQP for reranking search results. This would allow developers to easily toggle between 'diverse' and 'relevant' search results using a simple tunable parameter.

Suggested repo: nanoRerank

"Stop settling for redundant search results in your RAG pipeline."

Estimated effort: 40h