Design RAG Pipeline for PyTorch Optimization

Outline the architecture for your LlamaIndex RAG pipeline. What types of documents will you index (e.g., PyTorch docs, forum posts, internal code snippets)? How will you chunk and embed them? Describe how the agent will query this RAG system to retrieve relevant optimization strategies for a given PyTorch code problem.