Enterprise AI/RAG MVP Developer

Looking for a Python/AI developer to build a small working MVP for an enterprise AI platform. The MVP needs to demonstrate three things: 1. Automated Data Ingestion Use n8n + Google Drive to: Detect new/updated documents Extract and process content Apply Microsoft Presidio for PII/privacy protection Generate embeddings Store embeddings in Qdrant Make documents available for RAG Use existing tools and APIs wherever possible. Keep custom code to a minimum. 2. Chat-to-SQL Build a simple, safe natural-language-to-SQL capability. The MVP must work with: Wide World Importers — primary enterprise dataset Chinook — secondary test dataset PostgreSQL SQL Server Example: "Which customers generated the most revenue last year?" The system should generate and safely execute read-only SQL and return the answer. The LLM must not have unrestricted database access. Use a controlled database tool/skill with basic SQL validation, limits and logging. 3. Traceability Show what happened behind an answer using Langfuse or structured logs: User question → RAG/SQL → retrieved data → model → final answer. Preferred Stack Python, FastAPI, n8n, Qdrant, PostgreSQL, SQL Server, Presidio, LiteLLM, Langfuse, Docker. 4. Privacy & Data Redaction — Microsoft Presidio Use Microsoft Presidio to demonstrate privacy protection before data reaches an external LLM. The MVP should support: PII detection Redaction and/or pseudonymisation Configurable sensitive-data rules Processing documents/data before embedding or LLM processing Demonstration of what was detected and protected Privacy is a core part, not an optional feature. Deliverable A working, Docker-deployable MVP that can be demonstrated to a customer. This is not a research project. I want someone who can build a practical MVP quickly using existing tools wherever possible. Please send examples of similar LLM/RAG, Chat-to-SQL or n8n projects you have built.