colbert

Late interaction reaches cross-encoder-like quality with near-bi-encoder scalability. The signal is the per-token embeddings plus MaxSim matching that earns you the middle ground, and understanding what it costs. Here is the answer.

Late interaction keeps one vector per token, so a corpus that fit in a few GB as single vectors can swell 100x. Interviewers want to see you know the compression and indexing tricks (centroids, residuals, PLAID) that make multi-vector retrieval shippable.