Event
LA Hacks
Team
Puneet Bajaj
Rutuja Nemane
PhysioPal is a fully on-device iOS app built at LA Hacks 2026 that brings AI-powered physical therapy supervision to your iPhone. Built by Rutuja Nemane and Puneet Bajaj, PhysioPal won the ZETIC Company Challenge by tackling one of healthcare's most overlooked problems: patients doing physiotherapy exercises wrong at home, unsupervised, with no one watching.
What it solves
Millions of patients are prescribed physiotherapy exercises every year. Most do them wrong at home unsupervised. The supervision gap is the number one reason physical therapy fails. Continuous live monitoring is expensive and cloud-based AI solutions raise serious privacy concerns, especially for elderly patients. PhysioPal solves this without a single byte of patient data leaving the device.
What it does
PhysioPal puts a private, always-available physiotherapy companion directly on the device the patient already holds. It tracks your exercises in real time using your iPhone camera, counts reps, detects incorrect form, and gives corrective feedback. It reads your Apple Health data and adapts your routine automatically. Bad night's sleep? Deep squats become chair-assisted squats. If a fall risk is detected, the app calls your physiotherapist automatically and offers a live video follow-up.
How it works
PhysioPal uses ZETIC Melange to run pose estimation and an on-device LLM entirely on Apple Silicon with zero cloud inference. A Context Engine reads Apple Health data and adapts routines in real time. A Supervision Engine tracks the patient's skeleton via the iPhone camera using Melange SDK and Google MediaPipe. An Escalation Engine detects fall risk and automatically calls the physiotherapist via Twilio. No video, health data, or pose data ever leaves the phone.