About

Stephen Sweeney

AI governance architect · Technical writer · Monument, CO

A career that moved from network infrastructure through iOS development and software architecture arrived at AI governance by following one question: how do deterministic systems maintain authority when acting in probabilistic environments?

AgentVector is the formal answer. The writing on this site is the argument. Flightworks Systems is where it compiles to real hardware.

The Engineer Arc

The through-line is determinism.

Governed autonomy is the conclusion of a career that started with network infrastructure, moved through iOS development, and arrived at AI governance on Apple Silicon. Every chapter asked the same question in a different context.

2000 – 2008
Network engineer
LAMP stack on FreeBSD-derived Mac. Web hosting infrastructure. Determinism as first principle — DNS, routing tables, and firewall rules don't negotiate.
2008 – 2016
Software engineer
Apple ecosystem through the App Store era. Swift from day one. Actor isolation and value semantics as engineering discipline, not convention.
2016 – 2020
Software architect
The shift from "does it work" to "can we prove it worked, and that it will work the same way again." Auditability becomes a design requirement, not an afterthought.
2020 – 2025
Team lead
Architecture meets user workflow, measurable outcomes, and delivery sequencing. The PM artifacts are where governance becomes a product — and where you learn what actually ships.
2025 → now
AI governance architect
Building AgentVector, Flightworks Systems, and Chronicle Quest — using governed autonomy as both the product and the operating model. Writing publicly about what the building reveals.
Current Work

What's being built.

Governance Framework
AgentVector

A constitutional governance framework for autonomous AI systems. The Codex specifies what must be enforced. SwiftVector and RustVector enforce it. Domain jurisdictions apply it to aviation, desktop agents, and narrative AI.

The framework →
Autonomous Aviation
Flightworks Systems

FAA Part 107 operations with a governed AI layer. FlightGovernance applies AgentVector's Codex to autonomous drone operations — geofences, airspace authority, and replayable evidence on every mission.

flightworks.systems ↗
Narrative AI
Chronicle Quest

Human authorship protected as a formal governance constraint. ChronicleGovernance applies Laws 6 and 8 — Narrative Authority and Audit Integrity — to AI-assisted storytelling.

Field Operations →
This Site

Why agentincommand.ai.

This site exists because building AgentVector generated more ideas than could fit in implementation. The framework answers what governed autonomy requires and how to enforce it. The writing answers why — and why now, why here, and what the broader implications are for autonomous systems at scale.

Agency Papers are the position documents — the careful, citeable arguments. Agent Statecraft is the implementation layer — how you actually build this in production. Dispatches are the ongoing field notes from watching the AI governance conversation develop in real time.

For hiring teams

Technical writing portfolio.

Documentation samples are indexed at /portfolio — labeled by genre, audience, and what each demonstrates. The flagship sample is the Part 108 compliance scenario, written for regulators and compliance officers without prior framework knowledge.

For consulting engagements, see macsweeney.llc ↗.

Elsewhere

Find the work.