Platform Engineer - Compute Capacity

We are seeking a platform engineer to join our Compute Capacity team. This team owns the capacity plan that keeps compute supply ahead of demand across every region we operate in — the forecasting, buffer policy, and provisioning systems that make sure a project never runs into a wall it didn't know was there.

You'll work on the systems that turn a capacity plan into provisioned reality: reservation acquisition, fleet reconciliation, and the automation that keeps what we've committed to in sync with what we're actually running. You'll help build the metrics and alerting that let capacity problems surface months out, on vendor lead time, rather than at the moment someone needs the room.

You'll design, build, and operate systems that are both robust and highly automated — helping us hold the right buffer at the right cost, catch drift before it becomes a shortage, and give every team a single, trustworthy view of how much room we have across the millions of databases we manage.

• Help build and maintain the capacity plan that keeps Supabase's compute supply ahead of demand across regions and instance families
• Support buffer policy by modeling headroom targets and their cost tradeoffs for review and sign-off
• Build and maintain automation that turns the capacity plan into provisioned reality — reservation acquisition and top-up, fleet reconciliation, drift detection between committed and running capacity
• Extend our infrastructure as code for capacity-relevant provisioning
• Instrument capacity: build and maintain metrics for saturation, reservation coverage, idle buffer, forecast error, and provisioning latency
• Build and tune capacity alerting so headroom, quota, and reservation issues surface months out rather than at the moment of impact
• Support the demand gate — helping intake large customer commitments, launches, migrations, and new regions so they reach the plan before they reach the fleet
• Contribute to mitigation projects: right-sizing, instance-family migrations, autoscaling improvements, workload consolidation
• Help drive down compute cost per database through right-sizing, placement, and commitment coverage work
• Participate in capacity incident response and post-incident follow-through