Rag And Retrieval easy
Your team is switching to a newer, higher-quality embedding model. The existing 5 million chunks were embedded with the old model. What must you do?
- A Embed only new queries with the new model; existing document vectors remain compatible.
- B Re-embed the corpus into a new index, then switch queries over once it is complete.
- C Upsert only documents that change from now on, letting the index migrate gradually.
- D Project the old vectors into the new space by zero-padding them up to the new dimension.
Answer: B. Re-embed the corpus into a new index, then switch queries over once it is complete.
Different embedding models produce **incompatible vector spaces**, so query and document vectors must come from the same model; that forces a full re-embed, ideally built in parallel and then cut over. The tradeoff is a one-time compute cost in exchange for better retrieval quality. Option C is the right strategy for content changes, but here it would mix two vector spaces and make similarity scores meaningless.
One system design question a day, with spaced repetition so the ones you miss come back.
Get the appRelated questions
- You must index 100M chunks using a 3,072-dimension embedding in fp32, served from RAM via HNSW. The raw vectors alone take about 1.2 TB before graph overhead. You need to cut memory several-fold while losing at most a point or two of recall@10. Which approach fits?
- A multi-tenant SaaS app stores all tenants in one vector index with a tenant_id field. Queries fetch the top 20 chunks by similarity and then drop chunks from other tenants. Small tenants often get 0-2 results. Why does this happen, and what is the fix?
- Runbooks are split into fixed 300-token chunks with no overlap. When a procedure step starts near the end of one chunk and finishes in the next, answers come back incomplete. What is the lowest-cost fix?
- Your offline eval shows recall@5 of 92% on a labeled set, so the gold passage is in the context for nearly every question. Yet human graders mark 35% of answers wrong, and the wrong answers often ignore or contradict the gold passage. Where should the team invest next?
- Your product requires every RAG answer to include clickable source citations that users can verify. How should those citations be produced?
- A runbook RAG system uses fixed 500-token chunks with no overlap. Users who ask for a multi-step recovery procedure get answers that stop at step 4 of 9, and retrieval logs show the top hit ends mid-procedure. What is the most targeted fix?