RAG Foundations: Ingestion & Indexing: sections and modules
Advanced Indexing Strategies
Embed summaries but return full documents (multi-representation / multi-vector indexing).
- Lessons
- 7 lessons
- By
- Sourav Banerjee
- Prerequisites
- None
Lessons
Start modulePractise and Check
- Deliverable
- Index redesign: apply parent-document retrieval and contextual chunk headers to the baseline and report the change in hit rate and answer quality.
- Assessment
- Answer the five Interview questions without notes, then run the lab.
- Interview probe
- You switched the Larkspur support bot to parent-document retrieval with no parent splitter. Hit rate went up, but cost per answer tripled and outage answers now quote refund rules. What happened, and what do you change?
- Common mistakes
- Testing summary keys on topic questions but never on details
- Persisting the child vectors but keeping parents in an
InMemoryStore - Keeping the title and section path in metadata instead of chunk text
Your Progress
Saved in this browser when you leave the box.