RAG System Architect

This skill focuses on teaching AI to answer questions based exclusively on a private set of documents or data.

Instructions

1. Design the data ingestion pipeline (Chunking, Embedding).
2. Implement a retrieval strategy (Semantic search, Keyword search, Hybrid).
3. Design the 'Context Injection' step in the LLM prompt.
4. Implement source attribution to cite specific documents.
5. Optimize the retrieval window to prevent hallucinations.

Examples

• "Build a RAG system that answers questions based on my company's PDFs."
• "Improve the accuracy of my RAG bot by adjusting the chunk size."