Back to selected work
Independent product
PhotoRank AI
Independent Product Engineer·2026
An AI-powered image-ranking product with a mobile-first UX, vision-model analysis, and Stripe-based checkout — designed with a privacy-conscious, no-permanent-storage architecture.
The problem
Letting people rank and compare their own photos using AI analysis, without the product becoming a long-term store of their personal images.
What I owned
- Mobile-first Next.js application architecture
- Integration with OpenAI vision-capable models for configurable image analysis
- Stripe Checkout, payment verification and webhook handling
- Redis-compatible checkout ledger and temporary image-processing pipeline
Technical challenges
- Designing an image pipeline that supports AI analysis without persisting uploaded photos
- Verifying Stripe payments and webhooks reliably before releasing paid analysis results
- Keeping vision-model analysis configurable across different ranking packages
Engineering decisions
- Temporary, non-persistent image handling as a deliberate privacy constraint rather than a default backend pattern
- A Redis-compatible ledger to track checkout/entitlement state without a full database
- Mobile-first layout given the primary use case: uploading photos from a phone
Technology
Next.jsOpenAI vision modelsStripe CheckoutWebhooksRedis-compatible storage