Agentic work is full of uncertainty, and right now governance sits at the forefront — the field is still deciding how a population of autonomous agents should be specified, bounded, and held to account. That ambiguity is where I do my best work, because I've spent my whole career mapping systems and the triggers that move them.

The domains changed; the discipline never did. I studied Sociology and Psychology with a minor in Criminal Justice, and I specialized in mapping the societal and psychological triggers that push individuals toward extreme and serial violence — learning to read the conditions that precede a behavior rather than the behavior itself. From there I tracked safety under DOT regulations for a traffic safety company, working inside a body where law and compliance are not optional and the record has to hold. Then I assessed flow, risk, and operational analytics for an e-commerce company's warehouse, mapping how work moves, where it fails, and what a bottleneck costs downstream. Trigger analysis, regulatory rigor, operational flow and risk — all three converge on what I build now.

What I build is governance architecture for agentic AI: the roles, permissions, and approval logic that decide what an autonomous agent is permitted to do, separate from what it is capable of doing. My throughline is a single principle — competence is not authority — drawn from Divine Legality, my framework for reasoning about how legitimacy is produced, defended, and enforced inside a system. Authority isn't earned by being good at a task; it has to be designed, through role, mode, scoped permission, review gates, and escalation paths, or it doesn't exist.

That principle runs through two connected bodies of work. Me, Inc. is a personal AI operating system built as an accountable multi-agent organization, aligned to the NIST AI Risk Management Framework. The Tribunal — the research I'm bringing to Zenity — is its independent assurance body: a consolidated audit, security, and ethics unit of three arms where an item clears only if all three clear, the gate is tiered by risk and reversibility rather than seniority, and a human makes every dispositive call. The Tribunal recommends; it never rules. Its controls map onto EU AI Act obligations by design, and I state its limits — design-stage, uncalibrated thresholds, oversight that binds hardest for a lone principal — as openly as its contributions.

How I work: deliberately, but quickly. I architect the original frame first, then tinker with it as the work moves, checking that every layer aligns with the ones around it and that each is defined to the best of my capacity. I'd rather build the escalation path than the demo. I treat disclaimers, confidence labels, and approval records as load-bearing artifacts that carry verified meaning, not boilerplate to soften — and when I don't have the evidence, I say so and show the path to getting it.

I'll be direct about the thing that's true: I don't come from the same background as most people in this field. That's precisely why I want this audience to understand what I bring to the field. This is a passion, and I work diligently — constantly training myself on new systems, new data, new law, and new failure modes so I can be the strongest possible asset to any agentic team I'm part of. I map the triggers, I design the authority, and I keep a human at the center of every decision that matters.