alijafarkamal/Medical-Intelligence-Agent-Ghana

# MedDesert AI ### 🏆 1st Place — MIT Hack-Nation Global AI Hackathon (Databricks Track) **Hack Nation: Bridging Medical Deserts** An Intelligent Document Parsing (IDP) agent that extracts, verifies, and reasons over medical facility data from Ghana to identify medical deserts and infrastructure gaps. --- ## Architecture ``` User Question | v Streamlit UI ──> LangGraph Agent Graph ──> Databricks Free Edition (local) (local orchestration) (remote services) ``` **LangGraph** orchestrates a multi-agent graph. Each agent node calls **Databricks** services (Genie, Vector Search, Model Serving). **MLflow** traces every step. **Streamlit** renders the frontend. ### Agent Nodes | Node | Purpose | Databricks Service | |------|---------|--------------------| | **Supervisor** | Classifies intent, routes to correct agent | Model Serving (LLM) | | **SQL Agent** | Structured queries (counts, aggregations) | Genie (Text-to-SQL) | | **RAG Agent** | Semantic search over facility descriptions | Vector Search | | **IDP Extraction** | Extracts structured facts from free-form text | Model Serving (LLM) | | **Medical Reasoning** | Anomaly detection, plausibility checks | Model Serving (LLM) | | **Geospatial** | Distance queries, medical desert detection | Local (Haversine) + SQL | | **Synthesis** | Merges results into cited Markdown answer | Model Serving (LLM) | --- ## Prerequisites - **Python 3.13+** (managed via `pyenv` — see `.python-version`) - **Databricks Free Edition** account with: - A SQL Warehouse (auto-provisioned on Free Edition) - A Personal Access Token (PAT) - **Git** --- ## Quick Start ### 1. Clone and set up Python environment ```bash git clone github.com cd Hack-Nation # If using pyenv (recommended) pyenv install 3.13.4 pyenv local 3.13.4 # Create virtual environment python -m venv .venv source .venv/bin/activate # macOS/Linux # .venv\Scripts\activate # Windows pip install -r requirements.txt ``` …