Inspiration

Delayed MG diagnosis and real patient stories showed us that early warning can save function and quality of life.

What it does

MG-Care analyzes a 5-second voice sample to estimate fatigue risk, confidence, and clinical support insights.

How we built it

We built a voice-first pipeline with acoustic feature extraction, ML risk scoring, and a web app for recording, upload, and trend tracking.

Challenges we ran into

We had to handle variable audio quality, browser recording format issues, and reliable feature consistency across environments.

Accomplishments that we're proud of

We delivered an end-to-end prototype with live recording, AI-assisted outputs, and longitudinal history visualization in hackathon time.

What we learned

In healthcare AI, usability and clinical workflow fit are as critical as model performance.

What's next for Myasthenia Gravis (MG-Care)

We plan continuous monitoring and multimodal expansion with facial-analysis signals to improve early detection accuracy.

Share this project:

Updates