Phased AI Chatbot Development: Self-Hosted RAG System

Project: We seek an experienced AI/ML developer to build a self-hosted AI chatbot system utilizing Retrieval-Augmented Generation (RAG) for our health related course Website. We have a wiki, over 1,200 articles, a 70k Youtube channel, podcast, lots of content. End goal is to have everything relevant RAG-ed for LLM use.

We plan for users to have access to two tiers of LLM to answer their question:

Free Tier: Intelligent FAQ using public content, guiding users to resources.
Paid Tier: Expert-level assistant with secure access to proprietary support databases and premium materials.

**** This project is structured in paid, progressive milestones, beginning with a full prototype. We're looking for a long-term development partner. ******

Milestone 1: End-to-End RAG Chatbot Prototype (Pilot Phase)

Objective: Deliver a complete, functional RAG chatbot prototype. This milestone will serve as an immediate proof-of-concept, demonstrating core RAG pipeline functionality, response quality, and technical execution in a self-hosted environment.

Scope of Work:
Ingest and process a small sample of public website content provided.
Implement text chunking and generate embeddings using a local embedding model.
Store data in a local vector database (e.g., ChromaDB, FAISS).
Set up local LLM serving (e.g., Ollama) with open-source models (e.g., Mistral, Gemma).
Develop Python script: Query -> Retrieve Context -> Prompt LLM -> Generate Response.
Implement a simple Command-Line Interface (CLI) for interaction.

Deliverables for Milestone 1:
A fully functional Python script demonstrating the RAG process.
The populated local vector database files.
Clear setup instructions for local LLM serving and running the script.
A brief README explaining models and prototype functionality.
A demonstration of the chatbot responding to example queries.
Milestone 1 Budget: We have allocated $800 - $1800 USD for this pilot milestone. Please quote your price for this specific scope.

Evaluation Criteria: Functionality, AI response quality/relevance, code clarity & documentation, technical competence in local LLM/DB setup, communication.

Full Project Vision & Long-Term Income Potential:
Successful Milestone 1 completion leads to an invitation for subsequent phases, building the complete system. The estimated total development budget for the entire project is $7,000 - $20,000+ USD. Future milestones will involve: scaling data ingestion (public & proprietary), building a web UI, implementing secure paid-tier logic, and production deployment. This offers a significant opportunity for a stable, long-term engagement.

Preferred Technology Approach:
Python, LangChain/LlamaIndex, Ollama for local LLM serving, and local vector databases like ChromaDB/FAISS. Experience with FastAPI/Flask for backend APIs and Streamlit for UIs is a plus. We value documented expertise in open-source LLMs and RAG principles. We are open to well-reasoned alternative technologies that align with our self-hosted, cost-effective, performant goals.

How to Apply:
Please submit your proposal including:

Your quote for Milestone 1.
A brief outline (max 500 words) of your strategy for the full project's future milestones.
Your relevant experience with RAG systems, self-hosted LLMs, Python, and related frameworks.
Links to your portfolio or examples of similar AI development projects.