DevOps SRE

Kubernetes

-----------

Role summary

Provide frontline triage and resolution for Kubernetes cluster requests across AWS, Onprem AliCloud, and GCP. You will own tickets from intake through resolution in a Slack-based support channel, escalating to platform engineering only when a confirmed defect or capacity change is required. Singapore is the anchor site for this role, covering the APAC cluster estate at its point of origin.

Kubernetes core — must have

* Cluster lifecycle: create, update, delete, node pool management, stuck-state recovery

* Version upgrades and upgrade-failure diagnosis

* RBAC (ClusterRole / RoleBinding) and its side effects on node rotation

* kubeconfig and kubectl troubleshooting, API server behavior and request throttling

Multi-cloud — must have at least two

* AWS: EKS, Karpenter / EC2NodeClass, NLB and ingress-nginx, Route 53, subnet reconciliation

* AliCloud: ACK including region-specific Kubernetes version availability constraints

* GCP: GKE / GCE, vulnerability remediation workflows

* IS-Cloud and internal KCS control planes

Platform tooling — must have

* Kubernetes: cluster manifests, YAML schema validation, deployment error states, CLI

* Observability: Prometheus, kube-state-metrics, Splunk forwarding

* Add-ons: Spinnaker, KEDA, Compass compliance onboarding, backup services

Networking — must have

* Cross-environment DNS resolution and cluster-to-managed-database connectivity

* Load balancer reconcile loops and subnet annotation behavior

* Outage triage and formal RCA authorship

Required Qualifications

* 5+ years in Kubernetes operations, cloud infrastructure support, site reliability, or cloud operations engineering. The multi-provider scope sets this bar above a single-cloud equivalent role.

* Deep production experience operating managed Kubernetes on at least one provider (EKS, GKE, or ACK), with demonstrated working competence on a second. Depth in EKS or GKE is the most common qualifying profile.

* Cluster-level diagnostic depth: inspecting pod and node state, diagnosing volume and CNI failures, interpreting cluster events, and recognizing control plane and API server behavior under load.

* Strong identity and access management fundamentals that transfer across providers: role assumption and delegation, workload identity and service accounts, policy evaluation and troubleshooting, and how provider IAM binds to Kubernetes RBAC.

* Working knowledge of enterprise network fundamentals: routing, DNS, TLS, proxies, CIDR planning, and firewall/ACL models, plus how ingress and load balancer controllers reconcile against them.

* Scripting proficiency in Python or Bash, plus fluency with kubectl and provider CLIs.

* Ticket ownership in a shared support channel: intake, severity assessment, driving to resolution, and closing with a written outcome rather than a silent close.

* Excellent written English. The majority of support happens asynchronously in text, and clarity directly determines resolution speed.

* Legally authorized to work in Singapore.

Preferred Qualifications

* Certification in one or more of: CKA or CKAD, AWS Solutions Architect (Associate/Professional), Google Professional Cloud Architect, Alibaba Cloud ACP.

* Prior AliCloud or ACK experience, or demonstrated rapid ramp on an unfamiliar cloud provider. AliCloud talent is scarce in the market; evidence of learning a new provider quickly is an acceptable and expected substitute.

* Experience supporting internal developer platforms in a large enterprise.

* Familiarity with hybrid connectivity between corporate networks and public cloud, and with on-premise Kubernetes estates.

* Infrastructure-as-code exposure (Terraform, CloudFormation) and templating with Helm or Kustomize.

* Observability tooling experience (Splunk, Prometheus, Grafana, Datadog, CloudWatch, Cloud Logging).

* Prior follow-the-sun or multi-region support rotation experience.

* Familiarity with China-region cloud operations and their regulatory constraints.

* Track record of converting repeat escalations into runbooks or automation that measurably reduced ticket volume.