Hi PH — radiology AI scientist building HealthOS: voice biomarkers on iPhone

Hey Product Hunt — I'm Sabber. Long-time lurker, first-time poster.

My background is an unusual path to consumer health. I did a PhD in Geophysics (signal processing, Fourier transforms, spectrograms on seismic waves) then spent a decade shipping production AI — medical image segmentation in radiology at Sirona Medical, LLM agents for Medicare enrollment at Healthpilot, and real-time fraud detection at Asurion.

The through-line: turn messy real-world signals into something deployable and useful.

About 18 months ago I realized voice is the most underused health signal we carry. The same math I used on seismic waves applies directly to what your voice reveals — stress, fatigue, energy, breathing — all encoded before you consciously feel them.

So I built HealthOS: 30 seconds of unscripted speech, 8 biomarker reads, fully on-device. No wearable. Your audio never leaves your phone. Won 1st prize at Health Wildcatters' 2026 TXHCC Hackathon.

Pre-launch now with a TestFlight beta open. If you're into voice AI, on-device ML, or consumer health — I'd love to connect. Waitlist + beta: healthos.live