Proxy IA
AI generation platform using the Proxy pattern to enforce per-plan quotas and rate limits.
- Position
- Backend developer
- Role
- Backend and frontend development.
- Period
- Apr 2026
- Status
- Complete

Context
AI generation service with per-plan usage control.
The challenge
Prevent abuse and runaway costs on a paid AI service.
How I solved it
- Spring Boot API where a proxy intercepts every generation and checks quota and rate limit.
- Custom exceptions for exhausted quota and too many requests, plus a scheduled job that resets the rate limit every minute and quotas every month.
- React + Vite frontend; backend packaged with Docker.
Outcome
- The Proxy pattern keeps the business rule (quotas) apart from the model call.
Contact
Let's talk. Write to me or browse my code.