Storage Platform Engineer

Supabase · Remote, Global · Engineering

Posted 2026-11-19

Apply for this role →

ABOUT THE ROLE

Supabase runs millions of Postgres databases, and storage performance is now one of the biggest levers on how fast, reliable, and affordable they are. We're building a software-defined, NVMe-over-Fabrics storage platform underneath Supabase Postgres. Postgres stays unmodified, and the new layer delivers near-local-NVMe performance with the durability and flexibility of networked storage.

As the senior engineer driving storage, you'll own this effort end to end: the architecture and build-vs-buy decisions, integration into our compute platform, and operating the system in production.

Your first goal is raw Postgres I/O performance at scale. In the longer term, the layer you build will provide the block-level snapshots and copy-on-write clones behind database branching, fast restores, and instant resizes. Because it's software-defined, it will also travel with us as we expand beyond a single cloud provider into multi-cloud and colocated infrastructure.

This is a foundational, high-ownership role. You'll set the technical direction for storage at Supabase and help grow the team around it.

WHAT YOU’LL BE RESPONSIBLE FOR

- Lead the architecture of Supabase's NVMe-oF storage platform, including evaluating and deciding between adopting existing software-defined storage and building on SPDK

- Integrate networked block storage into our Kubernetes-based compute platform through CSI drivers and custom controllers, covering provisioning, attach/detach, resizing, and migration

- Make Postgres fast on networked storage by profiling and tuning the full I/O path, from fsync, WAL, and checkpoints through the kernel block layer, NVMe/TCP, and the network

- Design for durability and availability, including replication or erasure coding, failure domains, failover, and data-integrity verification

- Build I/O isolation and QoS so that noisy neighbors can't degrade other tenants

- Set performance and reliability SLOs, and build the benchmarking and observability to hold the system to them

- Design the layer for block-level snapshots and copy-on-write clones, laying the groundwork for database branching

- Keep the platform portable across AWS, other clouds, and bare-metal/colo

- Write design docs, mentor engineers, and work closely with the Compute, Branching, Postgres, and Multicloud teams

- Participate in on-call rotations, incident response, and post-incident reviews

- Resolve escalated support issues, while also identifying and resolving systemic issues across the platform

YOU MIGHT BE A GOOD FIT IF YOU

MUST HAVE:

- Have 7+ years of software engineering experience, including significant time building and operating storage systems in production

- Have built or operated distributed or software-defined block storage (for example Ceph RBD, Lightbits, Simplyblock, Portworx, LINSTOR/DRBD, or an in-house system)

- Have hands-on experience with NVMe/TCP or NVMe-oF, including targets, initiators, multipath, and failover behavior

- Have a deep understanding of Linux storage internals: the block layer, io_uring, page cache, O_DIRECT and fsync semantics, filesystems, and device-mapper

- Write strong systems code in Go, plus C, C++, or Rust

- Measure performance rigorously with fio, pgbench, eBPF, and perf, and can explain latency down to its individual components

- Understand durability tradeoffs between replication, erasure coding, and consistency under failure

- Have led technical direction on a complex system: design docs, tradeoffs, and bringing others along

- Have experience with Kubernetes controllers and/or CSI drivers

- Have been on call for large production systems

- Communicate clearly in writing and thrive in async, globally distributed teams

NICE TO HAVE:

- SPDK development experience

- RDMA/RoCE experience

- Experience implementing block-level snapshots, thin provisioning, or copy-on-write cloning

- Postgres internals knowledge (WAL, buffer manager, checkpointer), or familiarity with OrioleDB

- Experience with AWS networking and instance storage at scale (ENA, placement, bandwidth limits)

- Experience running storage on bare metal or in colo

- Contributions to open source storage projects or the Linux kernel storage stack

Apply for this role →

← Back to all jobs