Fitness trackers turn raw data into recovery scores, strain, and physiological trends — actionable insights, automatically. Hospitals have nothing like that. Clinicians in the ICU manually assess 100+ real-time measurements: ECG waveforms, blood pressure, oximetry, ventilation, perfusion. They're expected to derive insight from all of it, plus years of patient history and the details of the current stay.

It's overwhelming. And when a child's cardiovascular state can shift in minutes, manual monitoring isn't good enough.

So my team builds AI that does what fitness trackers do for athletes — but for clinicians caring for children who need specialty care, at far greater clinical depth and precision. Foundation models that read ECGs and generate explainable medical rationales. Mechanistic models that infer hidden cardiovascular states like contractility and vascular tone from waveform data. Changepoint detection that flags rhythm disruptions in real time.

All designed for transparency. All built to support clinical judgment, and undergoing rigorous validation. The goal is simple: give clinicians the situational awareness they need when seconds matter.