Lead DevOps Engineer

Requirements

Must-haves

• Experience leading or mentoring engineering teams, setting direction hands-on
• Strong Kubernetes knowledge: cluster lifecycle, API extensions, Operators, Helm, CNCF ecosystem (Cilium, ExternalDNS, Kyverno, Gatekeeper)
• Multi-cluster, multi-region Kubernetes platform design with Istio, Consul, or Linkerd
• Infrastructure-as-Code with Terraform on AWS or GCP, modular design, GitOps integration, automated testing
• GitOps pipelines with ArgoCD or FluxCD for progressive delivery and drift correction
• Containerised, serverless, or event-driven systems with strong observability (DataDog, Splunk, or OpenTelemetry)
• Vault-based secret management, least privilege access, compliance automation
• CI/CD workflows including SAST, DAST, policy enforcement, and performance telemetry
• Reliability and resilience through SLOs, error budgets, and chaos engineering
• Working knowledge of LLM-based services and AI infrastructure: deploying, securing, and operating AI Gateways and services
• Already built repeatable autonomous systems in production infrastructure settings, with a defined rule for what happens when they go wrong
• Taken a team through the shift to agentic or automated workflows and got it to stick

Nice-to-haves

• Platform modernisation or reliability initiatives in scale-up or regulated environments
• Operator development, CRD automation, eBPF, or Cilium for observability
• Policy-as-code using OPA or Kyverno within secure supply chain or CSPM frameworks
• Familiarity with MCP and A2A orchestration patterns in Kubernetes service mesh environments
• Agent Gateways and Registries connecting microservices and AI agents
• Secure containers, sandboxing, or confidential computing for regulated workloads
• Data-intensive systems such as Spark, Databricks, or Data Mesh
• Programming experience in Go, Python, or TypeScript
• Open-source or CNCF community contributions

What Success Looks Like

• A fully autonomous routine infra path is in production: agent writes the code, tests it, opens the PR, promotes through canary, and closes the loop without a human unless a policy gate or error budget trips
• Scaling, cert rotation, drift correction, and known alert patterns run without manual intervention
• The team's capability density around production agentic systems has measurably increased, with engineers able to build and operate autonomous workflows, not just use AI as a writing aid
• Guardrails are in place that define clearly where the autonomous line sits and what triggers human review, making it safe to move that line further over time
• Deployment velocity has increased without introducing regressions, drift, or policy violations

Team and Culture

• A technically strong team that already works at a high level, with high expectations of one another and no appetite for trial and error
• Opinions are expected to be backed by evidence and defended in front of peers who will push back, that is the normal mode of discourse, not an exception
• AI is already embedded in daily work across the team, and the direction is toward more autonomy, not less
• The culture rewards people who arrive with conviction and raise the bar, not those who wait to see which way the wind blows

Challenges

• The technical bar is already high, and this hire needs to arrive as the centre of gravity on agentic systems rather than growing into that position over time
• The environment does not tolerate a long runway of experimentation: decisions about where to move the autonomous line carry real consequences, and the reasoning behind them needs to hold up
• Building the guardrails that make production autonomy safe is as demanding as building the autonomy itself, and both need to happen in parallel
• The team needs to be brought up to production-level fluency on agentic workflows, which means carrying people through a genuine shift in how they work, not just introducing new tooling