I make infrastructure legible.

I came to cloud engineering the long way round, through support and operations, which is where I learned that most outages are really documentation failures. The estate worked; nobody could say why. That shaped how I build now.

My default is infrastructure as code. If a resource exists, it exists in a repository, with a plan you can read before it lands. Terraform for the estate, pipelines for the path to production, and enough observability that the first person to notice a problem is us rather than a customer.

I care about the unglamorous parts — tagging, cost allocation, least-privilege IAM, the runbook nobody wanted to write. They compound. A year of small, reviewable changes beats one heroic migration.