Senior Site Reliability Engineer​/SRE – Kubernetes & Hybrid Cloud; m​/f​/d

Introduction

At a glance

• Location &workmodel:
  
  Berlin, hybrid
• Tech stack:
  
  Kubernetes on our own servers, Harvester (Kube Virt), 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.

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.

• 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 Git Ops; 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

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 experienceisn'tenough for this role
• Lived SRE practice: SLOs, error budgets, incident management,on-call
• Hands‑on experience with
  
  GitOpsor comparable infrastructure/deployment automation– experience with Argo CD or Flux is a strong plus
• Solid observability skills– metrics, logs, traces, alerting that people trust
• A strong automation instinct– you'd rather fix a problem's cause than repeat its workaround
• A collaborative, enabling mindset– you see SRE as a service to our developers: you ask what they need, discuss trade-offs openly, and don't fall in love with your own solution

Nice-to-haves (genuinely optional –we'll teach you the rest):

• Harvester,Kube Virt, vSphere/ESXi, Open Stack or similar virtualization/HCI platforms
• Container storage (Longhorn, Ceph) and datacenter networking (load balancing, ingress, VLAN)
• Auto-scaling (HPA, VPA, KEDA, cluster auto scaler) and capacity/cost planning
• Experience…