Topic 1: Pushing Retrieval Quality

By the end of this module, you'll be able to:

• Build training data from real signals and fine-tune embeddings and rerankers, then prove the lift
• Run learned sparse and late-interaction retrieval at a size where their cost is real