Clipster
Backend, data model, and Discord integration for a creator clip-campaign platform. Built as the company's first engineer.
- Founding Engineer, Clipster
- Jan 2025 – present
- full-time, remote
- Go · Rust · TypeScript · PostgreSQL · Redis · Pub/Sub · Grafana
The problem
Clipster connects creators with paid brand clip campaigns. I joined as the first engineer and built the backend, the data model, and the Discord integration end to end. For roughly the first four months I was the only engineer; the team has since expanded. The parts that matter most are the ones with money on them: campaigns, creator earnings, referrals, and payouts.
The approach
- I built the backend in Go, with PostgreSQL as the system of record and Redis and Pub/Sub alongside it.
- I designed the data model with money in mind: earnings, referrals, and payouts have to reconcile, so every number a creator sees comes from queries over the same schema.
- I wrote the Discord bots in Rust. They handle account linking, DM delivery, notifications, and the rest of the platform’s presence in Discord.
- I keep the platform observable in Grafana.
The hardest part
Referral and payout correctness. Campaigns, referrals, and earnings all move money, often at the same time, and the numbers a creator sees have to stay right. I compute payouts and leaderboards from the source tables, with no separate copies to drift, so every figure holds up under concurrent writes.
Outcome
The platform runs in production at clipster.gg. Its own dashboards make it observable end to end.