Objective 3.5 · Domain 3 · 19% of the exam
Design a RAG pipeline with appropriate chunking and indexing strategies
Designing a retrieval-augmented generation pipeline: chunking, embedding, indexing, and the choices inside each. Chunk size trades recall against precision. Overlap protects meaning that straddles a boundary. Structure-aware splitting beats fixed-length splitting on documents that have structure. The most common wrong answer is a larger chunk size offered as a fix for poor retrieval quality, when the real problem is chunking that ignores the document's structure.
5 questions · answers and explanations shown as you go · free, no sign-up
The rest of domain 3: Integration
This domain is 19% of a scored form, about 12 of the 63 questions, spread across 8 objectives. Drill the whole domain or pick another objective below.
- 3.1Evaluate tool/agent configuration for capability bloat(5)
- 3.2Analyze authentication and authorization requirements to identify security gaps(5)
- 3.3Evaluate accuracy-latency trade-offs and justify configuration decisions(5)
- 3.4Analyze observability challenges and select monitoring strategies at scale(5)
- 3.6Apply retrieval strategies matched to data shape and query pattern(5)
- 3.7Evaluate connection protocols and select the appropriate integration mechanism (MCP, API/CLI, agent-to-agent)(4)
- 3.8Evaluate progressive discovery vs. monolithic context strategy(4)
One objective is a narrow slice. A full-length timed mock is what tells you whether the whole thing holds together.
Sit a timed mock →