Case study
AI-powered eyewear e-commerce
Production eyewear storefront and admin — Next.js commerce, Express/MongoDB API, AWS S3 media, Docker on Contabo, and browser-based AI virtual try-on with MediaPipe + TensorFlow.js.

Product overview — full frame
01 — Challenge
02 — Goals
03 — Signals
Measured strengths across delivery, innovation, and operational maturity.
28
Catalog models
On-device AI
Try-on innovation
Docker + SSL
Infra maturity
9 Redux slices
State architecture
04 — System design
Next.js frontend, Express REST API, MongoDB Atlas + S3 — each service containerized behind Nginx with SSL.
05 — Stack
Production tools powering this product — icons for quick recognition.
06 — Capabilities
Next.js App Router storefront + admin route groups
28 Mongoose commerce models
MediaPipe Face Landmarker try-on
Redux Toolkit + Zustand state
Docker Compose + Nginx HTTPS on Contabo
07 — Deep dive
| Layer | Tech |
|---|---|
| Face detection | MediaPipe Face Landmarker |
| ML runtime | TensorFlow.js + WebGL |
| Capture | React Webcam |
| Render | HTML Canvas (+ Three.js support) |
08 — Outcomes
Virtual try-on reduces purchase uncertainty
Unified admin for frames, lenses, and media
Containerized stack ready for growth
Closing note
A production-grade eyewear platform combining modern commerce UX with browser AI try-on and solid DevOps.