Astronomer empowers data teams to bring mission-critical software, analytics, and AI to life and is the company behind Astro, the industry-leading unified DataOps platform powered by Apache Airflow®. Astro accelerates building reliable data products that unlock insights, unleash AI value, and power data-driven applications. Trusted by more than 900 of the world’s leading enterprises, including Datadog, American Express, Expedia, Adobe, Autodesk, Together AI, and more, Astronomer lets businesses do more with their data. To learn more, visit www.astronomer.io.

About this role

Astronomer’s products run on a complex, multi-cloud platform — and the scalability, and operational maturity of that platform is what we’re actively investing in. The work ahead on our Platform team isn’t about managing data or curating pipelines; it’s about building the foundational systems that everything else runs on — the kind of systems where the failure modes matter, the latency budgets are real, and getting it wrong has visible consequences for hundreds of enterprise customers.

We’re looking for a Staff-level engineer who has built production platform systems at scale before — not just consumed them. You’ve designed systems under load, reasoned carefully about failure, and made your colleagues’ lives better by giving them something solid and well-understood to build on top of. You write strategy documents and then write the code that proves them out. You’ve driven a decision through, and lived with the consequences.

This is a foundational role at a consequential moment: your work will directly shape what Astronomer’s products — Astro, Observe, and our IDE — are capable of over the next several years.

What you get to do

Astronomer has a healthy and complex infrastructure estate spanning multiple cloud providers, a mix of managed and self-hosted systems, and an increasingly ambitious set of requirements as our products evolve. We have a clear sense of where we need to go; we need the right person to figure out how to get there and then go build it.

This is a deeply technical role with a high bar for code, so you’re an experienced developer, ideally in Go. A substantial part of your impact also comes from setting direction: design docs, proposals, and bringing along teams you don’t manage. Prefer heads-down coding most of the week? This role may not be the best fit.

Blaze a Trail: Own and develop our platform infrastructure strategy, with the sponsorship and responsibility to match. Map out what we need, build the case, bring the affected teams along, and own the outcomes.

Be an Owner: Be directly involved in deciding what we work on and how we work on it. Make promises, and keep them.

Do Sensible Things: Make principled build vs. buy assessments and advocate for the right tools for the right job — not the fashionable ones, not the ones already in the estate just because they’re there.

Garage Door Open: Create and maintain comprehensive internal documentation and decision records for systems and processes. Participate in architectural forums and make principled, open decisions that the rest of the organization can learn from and hold us to.

Is this role for you?

A small, senior team. No direct reports; you lead through influence, not management.

Broad scope. Our control plane, CI/CD, builds and releases, observability, secrets and networking, with internal teams as your customers.

Buy-in, not authority. You’ll often need agreement from teams you don’t manage.

What you bring to the role

Distributed systems depth, grounded in practice. You have a solid working model of how production systems fail — consistency and availability tradeoffs, failure cascades, backpressure, graceful degradation. You can draw the diagram, explain the failure modes at each node, and make a reasoned argument for which ones actually matter in a given context. NALSD thinking is how you naturally approach a new system design.

Kubernetes at operator depth. You know what happens inside the scheduler and the control loop when things go wrong, because you’ve been there. You’ve operated clusters under real load, not just deployed workloads onto them.