179 questions
No questions match those filters.
← All questions
Retrieval & EmbeddingsMedium· RRF· Retrieval fusion· System design
A teammate proposes running BM25 and dense retrieval in...
This is one of the questions in the full AI/ML interview bank. Pro unlocks all 1789 questions; Premium includes the same bank plus the highest daily Practice limit.
See plansReciprocal Rank Fusion (RRF) helps when one retriever misses what the other finds, but it does not shrink the input to the expensive reranking stages. If the corpus is large, both retrievers still search the entire corpus, which may exceed latency or compute budgets. RRF is a quality-improvement strategy, not a scaling strategy.