AngioSense

Remote and early screening for blood-flow issues is limited by access to specialists and diagnostic hardware.

An ESP32-based wearable streams impedance and SpO2 signals to a trained model that flags potential blockage, alongside a wireless stethoscope for remote auscultation.

Trained the classification model on body-vitals streams, designed the ESP32 firmware for wireless audio transmission, and built the companion React dashboard.

Getting a model to run reliably against noisy, real-world sensor data mattered more than squeezing out extra accuracy on clean training data.