How would you architect a retrieval system to solve this 'Multi-hop' problem? Example question: How does the Q3 revenue of our Tokyo office compare to the bonus of the CEO?

Question

Accepted Answer

Here are three ways to do this:

Query Decomposition: Instead of searching for the full string, you use an LLM to break the complex query into sub-questions.

Knowledge Graphs: If you build a knowledge graph of your data, you can traverse the edge between entities and find relationships that way

Agentic RAG: Tool use lets you use an LLM to decide what you need to answer and decide where and how to retrieve it

How would you architect a retrieval system to solve this 'Multi-hop' problem? Example question: How does the Q3 revenue of our Tokyo office compare to the bonus of the CEO?

Practice Your Response

Similar Questions in Context and Retrieval (RAG)