Performance Engineer - Performance Analysis & Tuning

Job description

About this role

Supabase is the Postgres development platform, built by developers for developers. We provide a complete backend solution including Database, Auth, Storage, Edge Functions, Realtime, and Vector Search. All services are deeply integrated and designed for growth. We're hiring experienced performance engineers to find and land performance wins across the Supabase stack — from the query optimiser to block device level, and across the network paths and cloud infrastructure that connect our distributed platform. This is a performance delivery role: real, measured improvements to our database, platform, and APIs as we scale toward enterprise and platform-grade workloads.

Responsibilities

• Find bottlenecks in live production systems and characterize them precisely enough that the owning team can act on them without you in the room.
• Partner closely with database teams (e.g. Multigres, OrioleDB) and infra teams to land concrete performance improvements.
• Build, communicate, and evolve performance methodologies and tooling that turn live production data into actionable insight.
• Partner with the observability team to capture the right signals and establish a unified view of platform performance (latency, throughput, tail behavior) across products.
• Help teams self-serve performance analysis and make performance a first-class engineering concern.
• Problems you might work on
• Tune the kernel's paging behavior to production workloads
• Reduce memory overhead in OrioleDB and Multigres
• Identify and eliminate sources of latency spikes from application to kernel level
• Identify scalability bottlenecks across the entire stack and work the appropriate teams in Supabase and the wider community to resolve them

Qualifications and requirements

• Deep experience in performance engineering, specifically performance analysis of complex distributed systems.
• You can walk the whole stack: follow a tail-latency problem from the query optimizer through the syscall boundary to the IO scheduler — or across pooler hops, network paths, and availability zones — reading profiles and traces at every layer.
• Depth on at least one end of that stack: Linux kernel internals (virtual memory, scheduling, networking, IO) or database internals (query planning, storage engines, buffer and WAL management)
• Strong working knowledge of networking and cloud infrastructure performance: TCP behavior under load, network virtualization overhead, and the performance characteristics of AWS primitives (EBS throughput and IOPS, instance network limits, placement groups, cross-AZ latency).
• Proven ability to turn performance insights into real product and architecture improvements — including getting teams you don't belong to to prioritize and ship the fix.
• Strong written communication: your problem characterisations are precise, reproducible, and compelling on their own.
• Pragmatic, tooling-oriented mindset; comfort working across product, infra, and database teams.