I'm a DevOps & Platform Engineer who builds for the worst-case scenario.

Own reliability and delivery for production Kubernetes across GCP, AWS and on-prem. Cut deployment failures 25% with blue-green Argo Rollouts and regression gates, migrated 300+ machines from CentOS 8 to Rocky Linux with a staged Ansible workflow, and standardised provisioning on Terraform across 150+ MySQL VMs and 10 GKE clusters. Built the agentic AI platform on LangGraph and Temporal that removed 1000+ manual hours a month. On the AIOps, MLOps and LLMOps side: AIOps self-healing across 175+ hosts that turns correlated Prometheus alerts into automated remediation, MLOps training pipelines on Argo Workflows with GPU node scheduling, and LLMOps observability with LangSmith tracing and per-workflow token and cost attribution.

• Kubernetes
• Argo Workflows
• GitOps
• Terraform
• AIOps
• MLOps
• LLMOps

Hardened Kubernetes workloads with PDBs, HPA, network policies and topology spread for a 40% availability gain. Moved releases onto GitOps CI/CD on ArgoCD for 50% fewer deployment failures and 40% faster deploys, cut container images 80% with multi-stage builds, and put Trivy, OWASP and Sealed Secrets in front of every release.

• Prometheus
• Grafana
• Tempo
• Thanos
• ArgoCD
• AWS

Fixes merged upstream into open-source projects including Argo CD (Flink health assessment), Prometheus node_exporter, prometheus-operator, Flux, MetalLB and Dapr. I contribute back to the tools I run in production, tracing bugs into upstream Go source rather than working around them.

• Go
• Argo CD
• Prometheus
• Flux
• MetalLB
• Dapr