Agentic-AI-Personal-Health-Coach

A digital health agent, powered by an LLM and deployed within each hospital or clinic, provides personalized recovery support to discharged patients. These agents are privacy-preserving, autonomous, and continuously learning across institutions using Federated Learning.

Business Problem Being Solved

• Current Challenges:
  • High readmission rates (especially for chronic conditions like heart disease, diabetes).
  • Lack of personalized care after discharge, leading to complications.
  • Patient non-adherence to medication or post-op care.
  • Data silos and privacy regulations prevent hospitals from sharing patient records.

Value Delivered

• Personalized, continuous care improves outcomes and patient satisfaction.
• FL ensures collaborative learning without violating privacy (no raw data sharing).
• AI agents automate repetitive tasks (follow-ups, reminders), saving nurse/doctor time.
• Hospitals reduce readmission penalties, and clinics can offer premium digital services.

Technical Architechture

• Core Components:

  • Local Agentic AI: An LLM-powered agent deployed in each hospital, acting as the digital health companion.
  • LLMs: Fine-tuned models like OpenChat, LLaMA, or GPT variants (on private infrastructure).
  • Prompt Engineering: Used to scaffold reasoning paths, guide conversation tone, and ensure medical safety.
  • Federated Learning: Model updates (gradients, not patient data) shared across institutions for collective intelligence.
  • Electronic Health Record (EHR) Integration: Pulls in clinical data with patient consent to contextualize recommendations.
  • Mobile/Web App: Patient interface for interaction, alerts, vitals input, reminders.

🚀 Installation & Running Instructions

Follow the steps below to set up and run the project locally:

1. Create a Python virtual environment

   python -m venv venv
   source venv/bin/activate   # On Mac/Linux
   venv\Scripts\activate      # On Windows
   

2. Clone the repository

   git clone https://github.com/ishant162/Agentic-AI-Personal-Health-Coach.git
   cd Agentic-AI-Personal-Health-Coach
   

3. Install dependencies

   pip install -r requirements.txt
   

4. Create a .env file Add your API key inside .env file in the project root:

   OPENAI_API_KEY=your_openai_api_key_here
    # or
   GROQ_API_KEY=your_groq_api_key_here
   

5. Run the app

   streamlit run app.py
   

Here's the user interface.