atominvest - Staff Engineer - Platform & Tooling

• BUILD THE MACHINE THAT BUILDS THE PRODUCT • Most of your time goes here. • Build our AI development harness. Self-service infrastructure, a reproducible development environment, and the sandboxes, guardrails and evals that let agents take on more of the work safely, with results we can measure. • Automate quality. Build the QA automation pipeline: automated tests, validation gates and release checks that give us confidence in every change, including the ones agents write. • Orchestrate agents across the business. Build pipelines that run agents reliably and safely in our product workflows and our internal implementation tools, not just in engineering. • OWN THE PLATFORM • Own our infrastructure. We're migrating to ECS. You'll strengthen our ECS architecture, see the move through in safe, reversible stages, bring every resource under infrastructure as code, and retire the old deployment model. • Build security into the platform. Tenant-scoped credentials, per-tenant secrets, least-privilege access, and the audit trail that SOC 2 and our clients' security teams expect. • Make production observable. Set up the monitoring, alerting and tracing we rely on, and build playbooks for on-call and incident response which is shared amongst the team. • WORK ACROSS THE STACK • Make the application changes the harness needs, such as synthetic data, test hooks and the specs agents work from. • Ship product. Pick up application work in TypeScript and React alongside the team. Over time, you'll learn the codebase as well as the platform it runs on. • Grow the team around you. Write the design notes and runbooks, pair with and review the work of other engineers, and mentor those who want to grow into infrastructure, so the platform never depends on one person. • YOUR FIRST SIX MONTHS • Get the first parts of "the machine" running, and show that they make the team faster without lowering quality. • At the moment all of our end to end testing is done manually by our Product team. That's a massive manual overhead that bottlenecks our product development. You'll help build our AI-powered QA automation pipeline, so every change, whether a person or an agent wrote it, is tested automatically before it reaches clients. • Strengthen the ECS architecture and complete the move with no disruption to clients. • Get the platform fully into code, with tenant isolation and least privilege designed in. • Put in place the monitoring and on-call foundations the team will build on. • HOW YOU'LL WORK • You'll own this area rather than work through a ticket queue. You'll decide what to build and when it's good enough, write your decisions down so others can challenge them, and tell us when something isn't worth doing. You'll work closely with Product from the start of significant work, because the platform exists to serve clients and the engineers building for them. • You and I will work together every day. We'll shape the technical direction together and think through the hard calls, and you'll have a direct line to every decision that affects your work. I'll back your judgement, and I'll expect you to push back when you think I'm wrong. • This is an individual contributor role, but we expect you to multiply the team: leading different areas at different times, and delivering through others as well as through your own work. • YOU'LL PROBABLY ENJOY THIS IF • You'd rather build a platform properly than keep patching one. • You're product-minded. You've probably worked at a startup or scale-up, or even founded something yourself. Seeing your work out in the world, solving real problems, is what motivates you. You ask who a change is for before you ask how to build it. • You already work this way. Agents write much of your code, and your time goes on design, specs and judging the output. You've built AI tooling for yourself, and probably for others on your team. • You get satisfaction from making other engineers faster. • You're comfortable writing Terraform, an API endpoint and an answer to a client's security questionnaire in the same week. • You've been the person others come to with hard problems, and you've changed your mind publicly when the evidence said so. • THIS PROBABLY ISN'T FOR YOU IF • You want a pure DevOps or SRE role. • You want to manage people right now. These opportunities might exist in the future, but for now we're small, lean and as flat as possible.