Practical agentic AI, from someone who ships it where the rules are strict

I'm an enterprise AI and process-automation architect working in regulated financial services. By day I design agentic systems that have to satisfy model-risk reviewers, auditors, and security teams before they satisfy anyone else. This site is where I write down what actually works — for anyone building with agents: students, hobbyists, startup builders, and enterprise teams alike. The regulated-industry experience is a bonus you'll feel in the depth, not a prerequisite for reading.

Three promises, so you know what you're getting:

• Everything is evergreen. No model-launch hot takes, no "top 10 AI tools this week." Tutorials and decision frameworks that will still be useful when you find them a year from now.
• Every code block runs. If it's published here, it was executed first — on real hardware, including ARM64. Troubleshooting sections contain errors I actually hit.
• The enterprise lens is real, and generic. I write about public regulatory concepts — SR 11-7-style model risk, OCC and FFIEC expectations, the EU AI Act — always genericized, never anyone's internals.

Where to start

New to agents? Start with the tutorial learning path. Choosing between approaches? Compare & Decide. Responsible for AI risk in a regulated shop? Enterprise & Governance is written for you.