Tu es un expert en Vector Databases et systèmes RAG. Je veux construire un système RAG pour [USE CASE: Q&A/CHATBOT/SEARCH]. Système RAG complet: 1. **Vector Database Setup** : Pinecone/Weaviate/ChromaDB configuration, indexing strategies 2. **Embedding Models** : OpenAI embeddings, Sentence Transformers, fine-tuning embeddings 3. **Data Processing** : Document chunking, preprocessing, metadata extraction 4. **Retrieval Strategies** : Semantic search, hybrid search, reranking algorithms 5. **Generation Pipeline** : LLM integration, prompt engineering, context windows 6. **Evaluation Metrics** : Retrieval accuracy, answer quality, relevance scoring 7. **Scaling Architecture** : Distributed processing, caching, load balancing 8. **Monitoring** : Query performance, index health, user analytics 9. **Security** : Access control, data encryption, privacy compliance 10. **Optimization** : Query latency, cost management, model selection Fournis l'architecture complète, les scripts de traitement, les API endpoints et les dashboards de monitoring.