Retrieval Quality & RAG Applications: sections and modules
Framework Fluency: LlamaIndex
Map LangChain concepts to LlamaIndex (Document/Node, Retriever, QueryEngine, NodePostprocessor).
- Lessons
- 6 lessons
- By
- Sourav Banerjee
- Prerequisites
- None
Lessons
Start modulePractise and Check
- Deliverable
- Port one of your T6 pipelines (hybrid + rerank) to LlamaIndex and confirm the metrics match.
- Assessment
- Answer the five Interview questions without notes, then run the lab.
- Interview probe
- You port Larkspur's hybrid retriever to LlamaIndex: a FAISS `IndexFlatL2` vector store and BM25, fused with `QueryFusionRetriever(mode="relative_score")`. The fused top result is now the least relevant handbook passage. What happened?
- Common mistakes
- Porting a character chunk size to
SentenceSplitter, which counts tokens - Feeding
IndexFlatL2distances into a fusion that assumes higher is better - Dropping cross-encoder nodes below 0, as if logits were probabilities
- Porting a character chunk size to
Your Progress
Saved in this browser when you leave the box.