Senior Site Reliability Engineer / SRE – Kubernetes & Hybrid Cloud (m/f/d)

Introduction At a glance Location & work model : Berlin, hybrid Tech stack: Kubernetes on our own servers, Harvester ( KubeVirt ), Argo CD/Flux, Prometheus/Grafana, Longhorn/Ceph Team: A growing SRE team – you report to our CTPO for now and to the Team Lead SRE we're hiring next; two system administrators in Pforzheim run the physical hardware Process: Intro call · take-home task (~2h) · 90-min tech interview with our developers · leadership conversation · meet the team Languages: Fluent English required; German is a plus, not a must Why this role is special Most SRE jobs today mean clicking around a managed cloud console. This one doesn't. We run our own hardware in Frankfurt and are building a modern private cloud platform on Kubernetes and Harvester – on-prem by default, with elastic burst into the public cloud and the option to go cloud-only later. You won't inherit a finished SRE practice: you'll help define it, side by side with our Berlin development teams – and you won't do it alone, a Team Lead SRE hire is coming next. SRE here is an enabling discipline: you build what our developers need to ship reliably, while two system administrators in Pforzheim run the physical hardware. And the impact is direct – our product discovery technology powers more than 2,000 European online shops (Intersport, SPAR, Douglas and more), handling billions of shopper queries a year. When product discovery is slow or down, our customers lose revenue in real time. Your first 90 days You get to know both products, join the on-call rotation with a buddy, and own your first reliability topic – SLOs for one product, alerting that actually helps at 3 a.m., or automating away a piece of toil. By day 90 you've shipped visible improvements and know where you want to take the platform next. Your mission Define and own SLOs, SLIs and error budgets; drive data-informed reliability decisions Lead incident response end-to-end: fast detection, clear communication, blameless postmortems – and reduce whole classes of incidents structurally, not case by case Eliminate toil through automation and GitOps; evolve our observability (metrics, logs, traces, alerting, runbooks) across two different stacks Help build our custom Kubernetes operator (CRDs) that makes stateful search clusters declarative, self-healing and safely upgradable – and roll out the auto-scaling (HPA/VPA, KEDA, cluster auto scaler) today's architecture makes hard Plan capacity, performance and cost across on-premises and cloud – including the large-catalogue and peak-season loads our merchants care about – and use AI tools wherever they measurably speed up diagnosis and operations Your profile Must-haves: Kubernetes in production – built, not just used : you've set up and maintained clusters on your own servers (e.g. kubeadm , RKE2, k3s) and know cluster lifecycle and upgrades – managed-only experience isn't enough for this role Lived SRE practice : SLOs, error budgets, incident management, on-call Hands-on e