RescuePulse AI: Proximity-Based Emergency Response

Prototype

🎯 Inspiration

In Switzerland, every second counts during a cardiac arrest. While ambulances take 8-12 minutes, local responders are often just 300m away. RescuePulse AI was built to bridge this "deadly gap" by connecting victims to trained citizens in seconds.

🚀 What it does

Smart Geofencing: Automatically alerts medical professionals within a 300m radius.
Dynamic Escalation: If no one accepts within 30 seconds, the radius intelligently expands to 400m and then 600m.
AI-Powered Guidance: Uses a RAG-based LLM to provide immediate, context-aware CPR instructions (e.g., adult vs. infant protocols) to the responder while they are en route.
Critical Data Sharing: Shares the victim's blood group and allergies instantly with the responder.

🛠️ How we built it

Backend: Python (FastAPI) for high-speed geospatial logic.
Algorithm: Haversine formula for real-time proximity calculations.
AI Agent: LangChain + Gemini for the medical guidance system.
Database: Firebase for real-time status updates and push notifications.

🧠 Challenges we faced

Optimizing the escalation logic to ensure low latency.
Designing a RAG pipeline that gives concise, life-saving instructions without "hallucinations."