Wound³ Inc. · Jan 2022 – Apr 2026

Wound³ software: iOS capture app, 3D engine, backend, and web app

The Wound³ product: an iOS capture app that turns iPhone depth data into measured 3D wound models, an on-device compute pipeline, a Python backend on AWS, and a React web app for clinicians.

My role: Built the iOS app end to end, much of the backend and web app, and parts of the on-device compute pipeline; managed the six-person software team.

Wound history view in the Wound³ web app (demo data).

3D reconstruction view, with the automatic wound mask and the volume, area, and depth computed from the model (demo data).

Clinician dashboard on an iPad, tracking a wound's volume, area, and depth across scans (demo data).

Scan review in the iOS app: automatic wound mask, 3D measurements, quality checks, and sync once back online (demo data).

Highlights

Wrote the entire iOS capture app in SwiftUI: guided capture with quality checks that reject a bad angle, distance, lighting, or focus, and a local-first design that keeps working without internet and syncs later.
Built parts of the on-device compute pipeline in Swift and C++ that turns iPhone depth data and camera intrinsics into measured 3D wound models with automatic wound masking.
Wrote much of the backend (Python, FastAPI on AWS with Postgres RDS and S3) and much of the React web app, and managed six developers across the 3D engine, backend, and frontend.

My contribution

I built the iOS app in SwiftUI end to end, much of the backend (Python and FastAPI on AWS, with Postgres on RDS and S3 storage), parts of the on-device compute pipeline in Swift and C++, and much of the React web app. I managed six developers across the 3D engine and R&D (three), backend (one) and frontend (two).

Technical details

The app converts iPhone depth data and camera intrinsics into measured 3D wound models; we compared stereo, LiDAR, and TrueDepth sensing across phone generations. Beyond measurement, the engine runs capture quality checks that reject a poor angle, distance, lighting, or focus, masks the wound automatically, and is local-first: capture works without connectivity, including in operating rooms, and syncs when a connection returns.

Tools and processes

Tools: SwiftUI; Swift; C++; Python; FastAPI; AWS (RDS Postgres, S3); React

Part of Wound³ Inc. (Fabri Sciences Inc.).