09 — Tools

Status: In progress. The eval & observability layer is written; vector database and framework comparisons are still planned.

File	What It Covers
01-eval-observability-comparison.md	Ragas vs. TruLens vs. DeepEval vs. LlamaIndex eval vs. LangChain eval — feature matrix, verified code snippets, decision criteria

What this section will cover

Opinionated, comparison-driven tool guides for each pipeline layer:

Vector databases (planned) — benchmark-backed comparison (Pinecone, Weaviate, Qdrant, Milvus, pgvector) with selection criteria beyond marketing pages
Frameworks (planned) — when LangChain/LlamaIndex/Haystack/LangGraph help and when they get in the way
Eval & observability (written — see Contents above) — Ragas, TruLens, DeepEval, LlamaIndex eval, LangChain eval compared on the same scenario. LangSmith/Arize Phoenix/Langfuse tracing tools are covered in 01_concepts/observability_and_evaluation_ops.md rather than duplicated here.

One comparison file per layer: feature matrix → decision criteria → migration notes → interview-relevant talking points.