Overview
What this challenge is about.
Build a Python pipeline that ingests raw PPG + accelerometer at 100Hz, applies motion-artifact rejection using the accelerometer channel, detects beats, computes RR-interval irregularity metrics (sample entropy, normalized RMSSD), and emits a binary 'likely-irregular' flag per 60-second window. Validate against the labeled cohort with patient-level sensitivity, specificity, and positive predictive value. Apply correct cohort splitting (no patient leakage between train and validation). Deliver the pipeline, a validation report with confusion matrices stratified by age and skin tone, and a 4-page limitations document for the medical-affairs team.
The Brief
What you'll do, and what you'll demonstrate.
Build a non-diagnostic AFib-detection pipeline from raw wearable PPG data and validate fairness across demographic strata without claiming medical intent.
Earning criteria — what you'll demonstrate
- Preprocess noisy wearable signals with motion-artifact rejection
- Implement beat detection and RR-interval irregularity metrics
- Evaluate biomedical models with patient-level (not sample-level) splits
- Communicate algorithm limitations to a regulatory-aware stakeholder
Program Fit
Where this fits in your program.
Sharpens the same skills your degree expects you to demonstrate.
Skills
Skills you'll demonstrate.
Each one shows up on your verified credential.
Careers
Roles this prepares you for.
Real titles. Real skill bridges. Pick the one closest to your trajectory.
Product Manager
Product managers in consumer health who have shipped a non-diagnostic feature understand the regulatory line and ship safely.
This challenge sharpens
- fairness-evaluation
- benchmarking
- health-sensing