Skip to content

README.md

Latest commit

 

History

History
184 lines (144 loc) · 5.2 KB

README.md

File metadata and controls

184 lines (144 loc) · 5.2 KB
Raw

🩺 NurseMate AI — Nursing Ausbildung Assistant

An AI-powered chatbot for nursing students starting their Pflegeausbildung in Germany. Built with a Retrieval-Augmented Generation (RAG) architecture to provide accurate, hallucination-free answers based on official nursing guidelines.

Developed over a dedicated 5-month learning period to demonstrate practical skills in AI/LLM integration and modern software development.

🚀 Live Demo

🛠️ Tech Stack

Layer Technology Why?
LLM Groq API (Llama 3.1 8B) Ultra-low latency inference
Orchestration LangChain + LCEL RAG pipeline management
Vector Store FAISS (local) No cloud limits, instant search
Embeddings HuggingFace all-MiniLM-L6-v2 Open-source, fast, accurate
Backend Django 5 + DRF REST API, scalable
Frontend React 19 + Vite 8 Modern, fast UI
Deployment Render (backend) + Vercel (frontend) Free hosting

📂 How It Works (RAG Pipeline)

User Question
     ↓
[FAISS] searches 2,500+ chunks of nursing knowledge
     ↓
[LangChain] builds context-aware prompt
     ↓
[Groq / Llama 3.1] generates answer based ONLY on retrieved context
     ↓
React frontend displays answer (no hallucinations)

📚 Knowledge Base Topics

  • ✅ Ausbildung structure, duration, and application process
  • ✅ Language requirements (B2/C1 German)
  • ✅ Salary during and after training
  • ✅ Vital signs measurement (blood pressure, pulse, SpO₂, temperature)
  • ✅ Hygiene and infection control (WHO 5 Moments)
  • ✅ Basic nursing care (Grundpflege)
  • ✅ Medication safety (5-R-Regel / 5 Rights)
  • ✅ Nursing documentation principles
  • ✅ Emergency recognition (CPR, stroke FAST, hypoglycaemia)
  • ✅ Recognition of foreign qualifications (Anerkennung)
  • ✅ Patient rights and medical ethics

🛠️ Local Setup

1. Clone the repository

git clone https://github.com/yourusername/nurse_chat_bpt.git
cd nurse_chat_bpt

2. Backend Setup

cd backend
python -m venv venv
source venv/bin/activate        # macOS/Linux
# venv\Scripts\activate         # Windows

pip install -r requirements.txt

# Set up your Groq API key
echo "GROQ_API_KEY=your_groq_api_key_here" > .env

# Build FAISS index
python ingest.py

# Run backend
python manage.py runserver

Backend runs at: http://localhost:8000

3. Frontend Setup

cd frontend
npm install

# Set API URL
echo "VITE_API_URL=http://127.0.0.1:8000/api" > .env

# Run frontend
npm run dev

Frontend runs at: http://localhost:5173

☁️ Deployment

Backend (Render)

  1. Push code to GitHub
  2. Go to render.com → New Web Service
  3. Connect GitHub repo
  4. Settings:
    • Root Directory: backend
    • Build Command: pip install -r requirements.txt && python manage.py collectstatic --no-input && python manage.py migrate && python ingest.py
    • Start Command: gunicorn core.wsgi:application --bind 0.0.0.0:$PORT
  5. Add environment variables:
    • SECRET_KEY (generate)
    • DEBUG=False
    • RENDER=True
    • GROQ_API_KEY=your_key
  6. Deploy!

Frontend (Vercel)

  1. Go to vercel.com → New Project
  2. Connect GitHub repo
  3. Settings:
    • Root Directory: frontend
    • Framework: Vite
  4. Add environment variable:
    • VITE_API_URL=https://your-backend.onrender.com/api
  5. Deploy!

📁 Project Structure

nurse_chat_bpt/
├── backend/                 # Django REST API
│   ├── core/               # Django project settings
│   ├── api/                # API endpoints
│   ├── rag_pipeline.py     # RAG engine (LangChain + FAISS + Groq)
│   ├── ingest.py           # Document ingestion script
│   ├── data/               # Knowledge base (txt + pdf)
│   ├── faiss_index/        # Generated FAISS vectors
│   ├── requirements.txt
│   ├── build.sh            # Render build script
│   └── manage.py
├── frontend/               # React + Vite
│   ├── src/
│   │   ├── App.jsx         # Main component
│   │   ├── api.js          # API client
│   │   └── components/     # UI components
│   ├── package.json
│   ├── vercel.json
│   └── vite.config.js
├── render.yaml             # Render deployment config
├── .env                    # API keys (gitignored)
├── .gitignore
└── README.md

🔒 Security Notes

  • The .env file is listed in .gitignore and will never be uploaded to GitHub
  • FAISS index is generated during build — no external database required
  • CORS is configured to allow only the frontend domain

👤 About

Built by Mustafa Şenoğlu and Müslüm Evin as a portfolio project demonstrating:

  • RAG (Retrieval-Augmented Generation) architecture
  • LLM integration (Groq / Llama 3.1)
  • Vector database operations (FAISS)
  • Full-stack development (Django + React)
  • Cloud deployment (Render + Vercel)