Project Summary
Implemented the referral attribution engine for Dropwash, allowing users to refer users and earn rewards. Architected to account for dynamic attribution parameters.
Context
Mobile app install journeys traditionally lose referral context during the App Store / Google Play install hop. To eliminate user friction and drive viral growth, I engineered a full-stack, zero-dependency Deferred Deep Linking & Attribution engine across Web (Next.js), Backend (NestJS/Redis/MongoDB), and Mobile (Expo/React Native) that seamlessly carries referral codes, promo discounts, and campaign links directly into the app on first cold start.
Key Contributions
- Cross-Platform Attribution Flow: Built a unified web-to-mobile capture pipeline that extracts multi-channel params (referrals, promos, UTMs, deep links) and auto-populates signup forms on the user's first launch.
- Anti-Collision Redis Resolution: Designed a dual-key caching strategy (
IP + Platform + AppType) with screen/OS metadata checks to ensure accurate device matching across shared cellular (CGNAT) networks. - Fraud & Rate Limit Protection: Implemented single-use cache invalidation upon resolution and strict IP rate limiting (5 req/min) to prevent replay attacks and referral bonus abuse.
- Automated Analytics & Reward Queues: Structured 90-day auto-expiring MongoDB lifecycle logs (
click→resolved) and integrated asynchronous BullMQ workers for transactional PostgreSQL wallet credits.