What I’m learning
Questions I am working through, with the projects and notes that made them concrete.
Current questions
How should an AI-assisted workflow fail safely?
I am studying where deterministic validation, human approval, and observable failure states belong around model decisions.
What makes a deployment workflow dependable?
I keep returning to the same concerns across cloud and business systems: repeatability, rollback, secrets, monitoring, and a clear owner when something fails.
Earlier project notes
Multi-region Kubernetes
Routing, failover, deployment packaging, and observability across AKS regions.
Security inside CI/CD
Adding vulnerability scanning to an AWS delivery pipeline instead of treating security as a final check.
Serverless workflow design
Using managed AWS services for authentication, data, orchestration, and order processing.