SkinPin: Cloud-Free On-Device Skin Condition Detector

#Healthcare

#Dermatology

Event

LA Hacks

Team

Rishvik Kambhampati, Samarth Hegde, Sankeerth Gandhari

SkinPin is a fully on-device AI skin analysis app built at LA Hacks 2026 that puts dermatology-level screening in everyone's pocket with zero cloud dependency. Built by Rishvik Kambhampati, Samarth Hegde and Sankeerth Gandhari, SkinPin tackles the reality that most people wait weeks for a dermatologist appointment for something that might be urgent.

What it solves

Skin conditions affect millions of people and timely detection matters. Most people have no way to get a quick assessment between doctor visits. Cloud-based skin analysis tools send sensitive photos to remote servers, raising serious privacy concerns. SkinPin puts dermatology-grade screening on the device already in your pocket, with no photos ever leaving your phone.

What it does

SkinPin classifies skin conditions across 30 plus dermatological categories from a single photo, from eczema and psoriasis to potentially malignant lesions. It gives a triage result of clear, monitor, see a doctor soon, or seek urgent care today. A body map lets users track conditions across different areas over time, with journaling and personalized insights all stored locally on the device.

How it works

SkinPin uses a two-stage model pipeline: a 16-class gate model detects normal skin versus any condition, and a 23-class disease classifier trained on the DermNet dataset handles specific condition identification. Both models run on-device via the ZETIC Melange SDK through a custom Swift native bridge built from scratch for React Native. A multi-agent triage pipeline combines image predictions with user-reported symptoms locally to produce a final recommendation with zero network calls.