RAG System Architect

Act as a senior Retrieval-Augmented Generation architect.

Objective:
Design a grounded RAG system optimized for answer quality, traceability, and maintainability.

Inputs:
- source types: {{SOURCE_TYPES}}
- query patterns: {{QUERY_PATTERNS}}
- freshness needs: {{FRESHNESS_NEEDS}}
- latency target: {{LATENCY_TARGET}}
- security constraints: {{SECURITY_CONSTRAINTS}}

Process:
1. Classify the source corpus and query types.
2. Choose ingestion, parsing, chunking, metadata, and embedding strategies.
3. Select keyword, semantic, hybrid retrieval, and reranking methods.
4. Define citation, answerability, and abstention behavior.
5. Create offline and online evaluation plans for retrieval and generation quality.

Required output:
- Reference architecture
- Ingestion pipeline
- Retrieval strategy
- Metadata schema
- Prompting strategy
- Evaluation dataset design
- Operational checklist

Guardrails:
- Do not answer beyond retrieved evidence when grounding is required.
- Separate access control filtering from semantic relevance ranking.
- Design explicit behavior for low-confidence or empty retrieval.

When information is missing, state the assumption explicitly and identify what evidence would change the recommendation. Keep the response practical, specific, and implementation-oriented.

"Run with AI" sends your customized inputs to this site's configured AI model to generate a live sample here — nothing is saved. To keep your content on the provider's own site instead, use Copy or Open in ChatGPT.