Jamil Rafique Shaikh
Senior SRE & Platform Engineer. 11 years in production - telecom, FinTech, cloud-native SaaS. Specialist in Talos Linux, multi-cluster Kubernetes, GitOps with ArgoCD & Flux, and AI-assisted reliability engineering. Available for contract engagements and consulting.
Migrating from Juju, RKE2, Rancher, or vanilla kubeadm to Talos Linux. I handle the full path: cluster design, node bootstrap, CNI configuration (Cilium + BGP), GitOps wiring with ArgoCD or Flux, and workload cutover with zero-downtime rollouts. Done this for multi-cluster setups with 5+ clusters.
Design and implement GitOps at scale - app-of-apps patterns, fleet management, environment promotion, tenant isolation, and policy enforcement with Kyverno. Works across ArgoCD and FluxCD. Practical delivery, not slide-deck architecture.
Full LGTM stack buildout (Loki, Grafana, Tempo, Mimir / VictoriaMetrics), Alloy agents for log and metrics export, SLI/SLO definition, alert routing, and on-call runbooks. I also build AI-assisted SRE agents that watch your cluster, detect issues, and propose fixes automatically.
Build the developer platform your team needs - without the Backstage overhead. Internal tooling, self-service deployment APIs, MCP server integrations for AI-driven deployments, secrets management with HC Vault, private image registries with Harbor, and GitOps-native everything.
- Medtronic - Lead SRE productionizing an industrial IoT manufacturing data pipeline for scale-out to 65 global production sites. Containerized and hardened a 4-component stack (HiveMQ MQTT broker, Ignition SCADA, PostgreSQL/CNPG, Kafka Connect) on Kubernetes Tanzu, delivering to a regulated production environment with a full Evidence Pack.
- Acquia - SRE for a leading enterprise digital experience platform. Established SLO-driven alerting, hardened Kubernetes deployment pipelines, and built incident runbooks that cut response time during on-call.
- Beamer - Improved production observability for a SaaS product used by thousands of teams: built Sumo Logic and Grafana dashboards, tightened deployment workflows, and introduced structured incident processes that reduced rollback frequency.
- Defined SLI/SLO thresholds and MTTD-focused dashboards across cluster health, storage I/O, and service availability - turning reactive firefighting into proactive reliability work.
- Turned CVE remediation from a manual quarterly chore into a tracked, GitOps-native operation: automated image update workflows, SLO-bound remediation timelines, and full audit trails across GKE/EKS clusters.
- EMTECH - Infrastructure and reliability engineering for central bank digital currency (CBDC) platform infrastructure. Zero tolerance for deployment failures; built guardrails to match.
- Built GitHub Actions and Azure DevOps pipelines with environment promotion gates, automated validation, and rollback triggers - eliminating the class of "it worked in staging" failures that plagued the team.
- Embedded SAST, SCA, and DAST into the delivery pipeline, reducing security incidents by 40%. Right-sized AKS workloads and tuned autoscaling policies, cutting cloud spend by 20% without touching SLAs.
- Built the CI/CD foundation for 5G Virtual Network Functions deployed for Samsung Korea - Jenkins pipelines, Helm charts with integrated tracing and alerting, and jMeter load validation at carrier scale.
- Led OpenShift cluster deployment on OpenStack, enabling containerized 5G workload operations that previously required manual, error-prone provisioning steps.
- Deployed production Kubernetes HA clusters on OpenStack for telecom operators across three countries - ePLDT (Philippines), Airtel (India), Lycamobile, and MTN Uganda. On-site delivery, not remote hand-holding.
- Wrote Ansible automation that replaced manual, undocumented deployment steps with repeatable, auditable playbooks - the kind of work that makes the next engineer's life easier instead of harder.
Container CVEs pile up when remediation is a manual process nobody owns. autofix-dojo watches DefectDojo for findings, maps them to running workloads, bumps image versions, opens PRs via GitHub/GitLab APIs, and tracks SLO compliance on remediation timelines - all within GitOps constraints. Prometheus metrics, web dashboard, zero direct kubectl access required.
Not a toy cluster. Talos Linux (immutable OS, zero SSH access), Cilium eBPF networking with BGP, ArgoCD App-of-Apps managing 27 live applications alongside FluxCD, Terraform-provisioned VMs on Proxmox, and Ansible bootstrapping - deployed end-to-end in 25 minutes. Running real products: a self-hosted PaaS (spinup.in), an AI SRE agent (openclaw), BPL production app, and Mattermost. Every pattern I deliver to clients runs here first.
An AI Site Reliability Engineer running as a cron-based agent on Kubernetes. Monitors cluster health, detects CrashLoopBackOff and pod failures, correlates incidents across namespaces, auto-deletes stuck pods within policy, and posts structured alerts to Mattermost. Phase 3 adds runbook intelligence and capacity forecasting. Powered by Claude Haiku to keep inference costs near zero.
Most K8s cost tools either require too much trust or produce useless recommendations. KubeWise connects read-only to your cluster, runs a recommendation engine against real usage data, estimates dollar savings, and generates patches you review before applying. FastAPI backend, Next.js dashboard, Helm chart for self-hosted deployment, and a demo mode pre-seeded with 20+ realistic workloads so you can explore without a live cluster.
Not a tutorial. A working push-to-deploy platform running on bare-metal Talos Kubernetes - the same thing Vercel sells, built from scratch. Go control plane, GitHub App webhooks with HMAC verification, Kaniko rootless image builds, per-tenant Postgres via CloudNativePG, Traefik routing, and Cloudflare external-dns automation. Multi-tenant, production-deployed, serving real traffic at spinup.in.