I keep clusters alive, clouds cheap and GPUs busy.

AWS → Azure, 100+ servers. Moved 5+ Kubernetes clusters and multiple TB of data across UAE, UK, US and India for 4 flagship SaaS products, including PostgreSQL, managed SQL Server and Lambda workloads. One Aurora PostgreSQL database alone was 186 GB with 411 schemas and 48,767 tables, moved via scripted dump/restore with row-count verification.

Platform. Containerized 4 Java/.NET products into 40+ microservices on AKS and EKS, managed through Rancher, with HPA, VPA, node autoscaling, Redis Sentinel HA and blue-green/canary releases.

Terraform + Helm. AKS/EKS, VNet/VPC, Key Vault, SQL MI/RDS and VMs as code; Elastic Stack, Grafana stack and app workloads via Helm.

Pipelines. 40+ microservices and 3 Flutter apps shipped through 3 TeamCity servers (250+ builds) and 8 GitHub Actions pipelines, covering 100% of services and apps. Replaced single-branch Git with a PR-driven QA → UAT → PROD promotion flow gated by SonarQube.

Observability & on-call. Prometheus, Grafana, Loki, Azure Monitor and CloudWatch with uptime SLOs, log retention on Blob/S3 and SES alerts for failed jobs. On-call for production incidents (CouchDB crashes, failed releases, version conflicts) — restored service and fixed root causes.

~$3,000/month saved. AI-generated monthly cost reports, reserved instances, monitoring-based rightsizing, unused-resource cleanup and S3 Glacier Deep Archive. Automated Azure Backup and AWS Backup across 6 regions for DR.

DevSecOps. Audited and consolidated 408 security groups across 9 AWS regions; rolled out Prowler, Gitleaks, SonarQube, Infisical and Cloudflare proxy.