Secure infrastructure · Zero-downtime deployments · Full observability
I design and operate production-grade AWS infrastructure — multi-AZ, zero-trust, fully automated. Every system I build is observable from day one, hardened by default, and deployable with a single command.
95%+ uptime on multi-AZ setups · 70% faster incident detection · 0 seconds downtime on live Kubernetes migrations · Zero SSH across all environments
| Cloud | AWS — EC2, VPC, IAM, ALB, S3, EKS, Lambda, CloudFormation, SSM, Secrets Manager, KMS, CloudTrail, SNS, Auto Scaling |
| IaC | Terraform (modules, remote state), CloudFormation (7-stack dependency chains) |
| Config mgmt | Ansible — dynamic aws_ec2 inventory, playbooks over SSM, OS hardening |
| Containers | Docker, Kubernetes EKS — Rolling Update, RBAC, ClusterRole, RoleBinding |
| CI/CD | GitHub Actions — Trivy scan, ECR push, OIDC/IRSA auth (zero stored credentials) |
| Observability | Prometheus, Grafana, CloudWatch Alarms, Log Insights, custom /metrics endpoint |
| Security | Least-privilege IAM, RBAC, KMS encryption, private subnets, SSM Patch Manager, UFW |
| Scripting | Python (Flask), Bash — automation scripts, health checks, cron jobs |
| OS & tools | Linux (Ubuntu 22.04), Git, Minikube, VS Code |
CloudFormationSSMAnsibleSecrets ManagerKMSCloudWatchSNSIAM
Full operations lifecycle — provisioning, security, config management, monitoring, DR, and incident response in one platform. 7 CloudFormation stacks in strict dependency order. Zero SSH — SSM only, every session logged to S3 and CloudTrail. Incidents simulated, RTO documented.
Terraform (Modular)EKSKubernetes RBACVPCIAM
Teams self-provision AWS environments by editing one config file — no tickets, no waiting. Separate ClusterRoles per environment (dev / staging / prod), composable Terraform modules, secure by default.
DockerEKSTerraformGitHub ActionsPrometheusGrafanaECROIDC
Live v1 → v2 migration with zero service interruption — Prometheus flat line is the proof. 5-stage CI/CD pipeline, maxUnavailable: 0 rolling update, OIDC auth — no AWS keys stored anywhere.
TerraformVPCEC2ALBIAMNAT GatewayBastion Host
95%+ uptime across 2 Availability Zones. Zero public EC2 exposure — private subnets, Bastion Host only. Full environment up or destroyed with one Terraform command in under 5 minutes.
CloudWatchPrometheusGrafanaPython FlaskSNS
70% faster problem detection — replaced manual log reading with automated CloudWatch alarms and Grafana dashboards. Custom Flask /metrics endpoint, Log Insights for root cause in seconds.
Helm · AWS SAA-C03 · GitOps / ArgoCD
