Retrieval Augmented Generation (RAG) starts and ends with embeddings. In between, we need to retrieve the text that is most relevant to the query from the user. This post is about this retrieval task