Dubai-based edtech platform built on Cloudflare Workers edge infrastructure. Cut costs by 60% versus VMs while delivering sub-120ms API latency at 5K RPS. SvelteKit frontend, GraphQL/Pothos API, PostgreSQL via Drizzle ORM, and Bunny.net video streaming.
Media
Key Contributions
- Edge Architecture: Architected edge microservices on Cloudflare Workers, cutting infrastructure costs by 60% versus VM deployments with near-zero cold starts at $400/month projected for 1M MAU.
- API Performance: Optimized GraphQL API to 120ms P95 latency at 5K RPS using Houdini client caching and Pothos persisted queries, reducing payloads by 45%.
- Notification Mesh: Designed multi-channel notification system orchestrating in-app (GraphQL Subscriptions, SSE, Durable Objects), push (VAPID/ECDH), and email (Postmark) with 99.9% delivery reliability.
- Search Engine: Built search and recommendation engine with PostgreSQL + Drizzle ORM using tag-weighted scoring, delivering sub-200ms results on 1M+ rows.
- Video Streaming: Implemented video infrastructure with Bunny.net HLS, custom Vidstack player, adaptive bitrate, and signed .m3u8 URLs, cutting buffering events by 80%.
- Session Management: Built session system extending Lucia v3 with custom Redis adapter (ioredis-Upstash), achieving sub-5ms lookups at $0.02 per 10K operations.
Skills
- Frontend: SvelteKit | TypeScript | Houdini (GraphQL) | Vidstack
- Backend: Cloudflare Workers | GraphQL (Pothos, Yoga) | Durable Objects | SSE
- Database: PostgreSQL | Drizzle ORM | Redis (Upstash)