Virtual try-on SaaS
Wear Instant
Private image handling and reliable asynchronous processing for virtual garment try-on.
In development
TypeScript
Workers
PostgreSQL
Overview
A virtual try-on product for ecommerce stores, combining a public presentation, merchant administration and backend generation workflows.
My contribution
Developed the backend, validated private image uploads, integrated Vertex AI and implemented idempotency, usage quotas, a transactional outbox and recovery logic, as described in the CV.
Product capabilities
- Private garment and shopper image handling
- Asynchronous generation jobs
- Usage limits and accounting
- Recovery and failure handling
Reliability across asynchronous boundaries
Idempotent requests, usage quotas and transactional outbox processing coordinate work across request and generation boundaries. Recovery logic handles failures rather than assuming every external generation succeeds.
Technical stack
Testing and reliability
Concurrency and failure scenarios were tested using PostgreSQL and Miniflare. The product remains in development, with the paid launch and Shopify App Store publication pending.