My team at HackPrinceton (Fall '25) built Rescura because no one should face health emergencies alone.
Our goal is simple yet powerful: to turn every phone into a digital first responder that recognises
signs of distress, guides lifesaving action, and calls for help when humans can’t.
Rescura was born from compassion and urgency, designed to make technology serve humanity where it’s
needed most.
Rescura detects early warning signs of cardiac emergencies and acts autonomously to protect lives.
Real-time monitoring: Analyses ECG and motion data to identify dangerous patterns such as arrhythmia,
immobility, or sudden impact.
Instant response: When risk is high, Rescura triggers SOS alerts with location data and begins calm,
step-by-step CPR coaching for bystanders.
Disaster resilience: During natural disasters or network outages, Rescura enters SOS Mode, continuing
to monitor offline devices and queue alerts until connectivity is restored.
Accessible safety: Designed for low-power, low-bandwidth environments, this technology brings intelligent
emergency response to places where it has never been before.
Rescura doesn’t diagnose; it empowers. It bridges the gap between danger and help, giving every person a
fighting chance.
On the technical side, we used signal processing, machine learning (LightGBM), and integrated Grok as our
AI agent, enabling real-time reasoning, response generation, and adaptive decision-making during emergencies.
Our stack included Python, React, ESP32, Arduino, and Chart.js, using data from the MIT-BIH ECG and WEDA-FALL
datasets.
I converted our fall data to 125Hz format from 5 to 50Hz data and generated synthetic data while incorporating
error handling into all methods. I worked on fine-tuning LGBM parameters. I collaborated on pitch development
and presentation.