Writing
Field notes and essays on security, AI agents, technology leadership, and the infrastructure I run.
Browse long-form series by topic →
Start here
-
Why I Run Production-Grade Personal Infrastructure
Running personal systems with production discipline exposes blind spots that professional environments can hide.
-
Most of Building an Agent Isn't the Agent
The difficult parts of building a personal agent are identity, integration, deployment, and operations - not the model call.
-
Layered Terraform on GCP
Operating a personal system across independent Terraform layers, with separate state, scoped IAM, and explicit secret boundaries.
-
The Privileges I Gave Away: When Agentic AI Meets Your Trust Boundaries
What happened when I gave an AI agent access across systems whose trust boundaries were never designed for it.
-
From Vibes to Production: Governing AI-Assisted Development for Knowledge Workers
A practical architecture for governing applications built by knowledge workers with AI assistance.