The third piece in our in-depth series picks up where part two left off—with vector and lexical scores side by side, deliberately unfused.
It shows how the agent retrieves from it: filter first, generate vector and lexical candidates, fuse and rerank them, then budget the
