The RAG Interview

Book index

This Notion edition links forty-one chapters, nine front-matter units and seven appendices: $41+9+7=57$ content units. This start page, Reading Routes and Source Index are three additional navigation pages, making sixty indexed pages. Links open Notion pages; the titles and source page labels do not imply downloadable Markdown files or current Notion page numbers. Use the route guide for selected study and the source index for original-edition spans and declared prose-only reading estimates. 01-READING-ROUTES · 98-SOURCE-INDEX

Roadmap

Part	Chapters	Focus
Front matter	$00.01$ – $00.09$	You learn the scope, study method, notation, and intended use of the edition.
I. The RAG Interview Landscape	1–3	You frame the problem, compare alternatives, and structure a design answer.
II. The Generator Side	4–8	You examine generator capabilities, possible failure causes and model-capacity economics without assuming one unique owner for every error.
III. Prompting and Context Construction	9–11	You control instructions and assess prompt sensitivity, abstention, and calibration.
IV. Representing What You Retrieve	12–14	You choose representations, retrieval units, and ways to preserve document structure.
V. Indexing and Vector Search	15–17	You choose and operate an index within recall, latency, memory, and update budgets.
VI. Retrieval and Ranking	18–23	You compare retrieval families and ranking stages by quality and serving cost.
VII. Query Understanding and Control Flow	24–26	You reformulate queries, route requests, and control repeated retrieval.
VIII. Training the RAG System	27–29	You decide what to train and how to obtain useful supervision.
IX. Generation and Context Assembly	30–31	You manage context position and check the evidence behind generated claims.
X. Evaluation	32–34	You measure retrieval and generation separately and together, then test regression hypotheses and interactions.
XI. Trust, Credibility, and Adversarial Robustness	35–36	You evaluate credibility signals, provenance and access boundaries without treating authenticity as proof of truth or a security guarantee.
XII. Scaling, Advanced Variants, and Design Drills	37–41	You apply the earlier decisions to scale, federation, multimodality, graphs, and design drills.
Appendices	A1–A7	You look up formulas, test yourself, review designs, and prepare system documentation.

Front matter

Part I. The RAG Interview Landscape

Part II. The Generator Side

Part III. Prompting and Context Construction

Part IV. Representing What You Retrieve

Part V. Indexing and Vector Search

Part VI. Retrieval and Ranking

Part VII. Query Understanding and Control Flow

Part VIII. Training the RAG System

Part IX. Generation and Context Assembly

Part X. Evaluation

Part XI. Trust, Credibility, and Adversarial Robustness

Part XII. Scaling, Advanced Variants, and Design Drills

Appendices