AI/LLM Engineer for RAG-Based SaaS Platform with React, Python & AWS

We have an existing React-based SaaS platform where users upload internal documents and need to ask questions against that content. Our current OpenAI integration works for basic prompts, but retrieval becomes unreliable as the document set grows—responses occasionally miss relevant context, pull from the wrong documents, and take too long to return. We need an AI/LLM engineer to fix this by building a proper Python/FastAPI RAG pipeline. The immediate work is to handle document parsing/chunking, embeddings, vector retrieval with tenant/document-level filtering, and grounded OpenAI responses with source references. The RAG service then needs to be connected cleanly to our existing React application. The final piece is production deployment on AWS using Docker, including secure API access, error handling, logging, and retrieval/latency optimization. The goal is specific: users should be able to upload their company knowledge and receive fast, accurate answers based only on the content they are authorized to access.