Sohan Dogra
Platform & DevOps Engineer
Faridabad, India
I build reliable cloud infrastructure, automate delivery, and run production Kubernetes platforms. Four years of GitLab CI/CD at scale — runner fleets tuned at the OS level, reusable pipeline templates other teams build on, and security gates that run on every commit. Hands-on with AWS, EKS, Terraform and Argo CD GitOps, including production cluster upgrades and multi-account governance.
Delivery pipeline I run
Commit to production, with the gates that run on the way.
- GitDesired state
- GitLab CIBuild · test
- Security gatesSAST · Gitleaks · Trivy
- RegistrySigned images
- Argo CDApp-of-Apps sync
- EKSIstio canary
- ObservabilityPrometheus · Grafana
Rollback is a git revert — the same path in reverse, not a separate runbook.
How I engineer
Infrastructure as Code
If it was clicked in a console, it will drift. Terraform modules and Ansible roles with real test coverage, applied through pipelines rather than laptops.
Security by default
Scanning belongs in the shared pipeline stage every team already consumes, not in a checklist someone remembers. It runs on every commit or it does not run.
GitOps
Git is the desired state. Promotion between environments should be a declaration, not a runbook — which also makes rollback a revert.
Reversible rollouts
Canary and progressive delivery exist so production changes have a middle state. An upgrade plan without a rollback plan is not a plan.
Observability before incidents
Dashboards and alerting for pipeline and platform health, so failure trends are visible before someone escalates them.
Self-service over tickets
The platform team should not sit in the path of every deployment. Reusable templates and App-of-Apps let product teams move without waiting.