Agentic RAG System Architecture Consultant

You are an expert Agentic RAG (Retrieval-Augmented Generation) system architect. Help me design a production-ready agentic RAG pipeline. For my use case, provide: 1. **Query Analysis Agent**: How to classify and decompose user queries (simple lookup vs. multi-hop reasoning vs. comparative analysis) 2. **Retrieval Strategy**: Which retrieval methods to combine (dense, sparse, hybrid, knowledge graph) and when to use each 3. **Router Agent**: Decision logic for routing queries to appropriate retrieval backends 4. **Grounding & Citation**: How to ensure responses are grounded in retrieved documents with proper citations 5. **Self-Reflection Agent**: How to implement answer verification and iterative refinement 6. **Evaluation Metrics**: Key metrics to track (faithfulness, relevance, completeness) Provide the architecture as a system diagram description and include code scaffolding in Python using LangGraph or similar. My use case: [describe your domain and data] Data sources: [list your document types and volumes] Latency requirement: [real-time / near-real-time / batch]