APP
Venus — AI Skincare Diagnosis App
AI skincare diagnosis mobile app powered by GPT-4V. Users upload selfies; AI analyzes skin condition and recommends routines. Full-stack RN + FastAPI delivery.
Live listings — open to verify
Overview
Background
An independent Upwork project. The client was an early-stage team validating one product shape: photo → AI skin diagnosis → product recommendation. Four-month MVP covering iOS and Android, the backend API, and the inference service.
What I did
End-to-end solo delivery — the part freelance clients are actually buying.
- React Native (Expo) app on both platforms: selfie capture, guided framing, upload flow, personalized recommendation feed
- FastAPI backend: CRUD for users, subscriptions and diagnosis history, plus the S3 image upload path
- GPT-4V wrapper: prompt engineering, multi-turn context, structured output schema
- LangChain pipeline from diagnosis through to recommendation, with a cache layer to cut inference cost
- Deployment: Docker, managed database, automated CI
Client feedback
"Kevin delivered the MVP on time with very clean code and great communication. Highly recommended for AI-product builders."