Staff Backend Engineer

About Kestrel Labs

We ingest a few million spans a second on a normal Tuesday and rather more when one of our customers has a bad day. Kestrel Labs is a remote-first company of about 130 people building the tracing and metrics backend that sits underneath all of that. Engineering is organised into small squads with a lot of autonomy and a strong bias toward writing things down.

What you'll do

• Take technical ownership of the ingest pipeline: the path from an agent on a customer's machine through to queryable, retained telemetry.
• Lead the work to move our storage tier onto a columnar format, including the migration plan, the dual-write period, and the rollback story if it goes badly.
• Set the direction on sampling and cardinality control, which is currently the largest driver of both our infrastructure bill and customer complaints.
• Write designs that other engineers can disagree with productively, and shepherd them through review to a shipped result.
• Improve how we run in production: better alerting on our own systems, meaningful SLOs, and incident reviews that change something.
• Mentor senior engineers on the squad, mostly through code review and design feedback rather than formal management.

What we're looking for

• Eight or more years building backend systems, including several years on something with genuinely difficult scale or latency constraints.
• Deep experience with Go, or enough depth in a comparable systems language that picking up Go is a non-event.
• A working understanding of storage engines: how data is laid out on disk, why compaction matters, and where the sharp edges are in your chosen database.
• Operational maturity. You have been on call for something important and have opinions about what makes that sustainable.
• The ability to make a case in writing and change your mind in public when someone makes a better one.
• Comfort working asynchronously across a wide spread of time zones.

Nice to have

• Exposure to OpenTelemetry, Prometheus or similar telemetry ecosystems.
• Experience running multi-tenant infrastructure where one customer's load must not become everyone's problem.
• Prior work on a large migration that was completed rather than abandoned.

Benefits

• Health insurance for you and your dependents, or a cash equivalent where we cannot contract locally.
• Meaningful equity with a ten-year exercise window.
• USD 2,000 to set up your workspace plus a hardware refresh every three years.
• Genuinely async: no standing daily meetings, and core collaboration happens in writing.
• USD 2,500 annual learning budget.