Six years running the HMI/SCADA systems of a thermal power plant at 99.9% availability. Then industrial cybersecurity under IEC 62443. Today, the same discipline applied to cloud infrastructure: pipelines, observability, and production that cannot go down.
Four fronts. In all of them the deliverable is the same: someone who answers when the system stops at three in the morning.
Assessment, landing zone design, and execution. A real inventory before the plan, defined waves, and a rollback runbook written before the cutover. FinOps included, because migrating without controlling cost only moves the problem.
Terraform as the default way to build, pipelines in Azure DevOps, GitHub Actions or Jenkins, and Kubernetes when the problem actually warrants Kubernetes. Environments that rebuild from scratch instead of getting patched.
Prometheus, Grafana, Loki, Azure Monitor. The hard part is not the tooling, it is the alerting strategy: alarms that map to user-visible impact, not to resource thresholds. An on-call rotation that receives noise stops reading the pager.
Identity and secrets management, infrastructure hardening as code, and compliance where it carries weight: LGPD, ISO 27001, SOX, HIPAA, and IEC 62443 in industrial environments.
A schematic record, drawn in the format of chart recorder paper. The trace follows the criticality of the systems that became my responsibility; the four marked points are the turns that explain the path.
Not a list you have to take on trust. Microsoft publishes a transcript anyone can open, and the AWS validation number is below.
The questions that usually cost two or three emails, answered here instead.
The first conversation is a diagnostic, not a sales pitch. Thirty minutes, no charge. If I am not the right person for your problem, I will say so on the same call.
