Mohamed Salah — 8+ years turning traditional infrastructure into scalable, secure, cloud-native platforms across AWS, Azure and GCP. Cut cloud spend 40%+ migrating 50+ microservices to Kubernetes, unified 8 clusters onto one observability platform, and kept CPaaS traffic at 99.95% availability. Writing devops.msrashed.com to share what actually holds up in production.
Every reliability conversation starts with an SLI. I've run the DORA metrics, SLO, and error-budget practice behind an 8-cluster fleet — opinions are cheap, a burn-rate chart ends the debate.
DevSecOps end to end: image and IaC scanning, admission control, runtime policy enforcement, Vault and Keycloak for secrets and identity — built in from the first commit, not patched on before an audit.
Terraform and ArgoCD-driven GitOps replaced manual promotion across 50+ services. If a human does the exact same thing every time, that's a bug in the process, not a task for a human.
Own GKE clusters end to end across AWS, Azure and Google Cloud; architected agentic AI operations with a dedicated Veirox instance and custom MCP servers for Prometheus, Loki, Tempo and cluster state. Mentor a team of 3 DevOps engineers.
Migrated 50+ microservices from EC2 to ECS and then to a cloud-agnostic Kubernetes estate, cutting cloud spend by 40%+. Built the central observability platform unifying 8 clusters (EKS, GKE, AKS, on-prem Rancher), cutting MTTR from hours to minutes.
Architected multi-tenant, multi-zone AWS services across EU, US and India for a CPaaS platform handling millions of daily SMS and voice transactions, sustaining 99.95% availability.
Managed Mobile and Internet Banking deployments across on-premise and AWS in a regulated hybrid environment, tuning performance against banking SLAs.
Developed and optimised the Vodafone Egypt Payment Gateway, tuning performance across MySQL, PostgreSQL and Oracle.