N8N AI Automation - RAG, Vector Database Experience Required

We are looking for an experienced AI Engineer to build an intelligent RAG (Retrieval-Augmented Generation) chatbot capable of answering customer questions based on a knowledge base generated from Nike Store related Docs and or PDF documents. The system should ingest one or more complex Docs/PDF documents, parse and chunk the content, generate embeddings, store them in a vector database, and use an LLM to provide accurate, context-aware responses. The chatbot should also perform calculations and answer questions derived from the information contained within the PDFs. This is not a generic ChatGPT integration—we require a production-ready AI solution with retrieval, reasoning, and structured workflows. Functional Requirements: - Answer questions based only on our Nike brand related complex docs and Pdf. - Search semantically rather than by keyword. - Retrieve the most relevant document sections. - Cite or reference the relevant document section where feasible. - Handle follow-up questions using conversation history. - Perform calculations derived from document data. - Explain pricing, discounts, policies, or specifications if contained in the PDFs. - Respond naturally while remaining grounded in the retrieved content. AI / LLM - OpenAI GPT-4.x / GPT-5 (or Claude) - RAG architecture - Prompt Engineering - Embedding models - Function / Tool Calling Vector Database (experience with one or more) - Pinecone - pgvector - Qdrant - ChromaDB