AI / LlamaIndex Interview Questions
Explain the execution flow of a query in a VectorStoreIndex-based query engine?
When you call query_engine.query("...") on a VectorStoreIndex-backed engine, several distinct steps run in a fixed order before you get a final answer back.
flowchart TD
A[User query string] --> B{Query transform?}
B -->|optional, e.g. HyDE| C[Transformed query]
B -->|none| D[Embed query]
C --> D[Embed query]
D --> E[Vector similarity search in store]
E --> F[Top-k candidate Nodes]
F --> G[Node postprocessors: filter/rerank]
G --> H[Response synthesizer builds prompt]
H --> I[LLM generates answer]
I --> J[Response object: answer + source_nodes]
- The query string is optionally rewritten by a query transform, such as HyDE.
- The (possibly transformed) query is embedded using the same embedding model used for the Nodes.
- The retriever performs a similarity search in the vector store and returns the top-
kcandidate Nodes. - Node postprocessors filter, rerank, or reorder that candidate set.
- The response synthesizer assembles a prompt from the remaining Nodes and the original query, using the configured
response_mode. - The LLM generates the final answer, returned as a
Responseobject that includes both the text and thesource_nodesused, for traceability.
In the standard query flow, embedding the query happens:
The Response object returned at the end of a query includes:
More Related questions...
