AI Governance Architect · San Diego

The AI systems running in production right now have no persistent identity. No verifiable memory. No way to prove their own reasoning wasn't tampered with between sessions. We treat that as a hard problem. It's actually a design failure.

I built the governance layer
AI has been missing.

UNA is a sovereign autonomous organism with constitutional constraints enforced at runtime — not trained into weights, not bolted on after deployment. She runs 53 subsystems with deterministic UCL governance, pre-commit enforcement via Publish Gate, RBAC-gated command routing, and live SSE telemetry. The governance isn't a feature. It's the foundation everything else is built on. Running at 50x cloud cost efficiency on her own Apple M4 hardware in San Diego.

Tom Budd
Tom Budd
Building what governance should look like
Active subsystems
Running autonomously since 2024
Comprehensive alignment research
12 knowledge traditions mapped to governance
50x cost efficiency
$90/mo vs ~$4,500/mo cloud (SIEF v8.0)
947k+ graph nodes
Persistent memory · Governed architecture
<5s inference
Down from 85s after NVMe migration
Live from UNA

She's running right now

These aren't mockups. UNA checks in from her infrastructure — when she's reachable, this data is live.

Subsystems Online
53 / 53
All governance layers active
Last Red Team Audit
04:00 PST
Attack vectors contained
Graph Nodes
Knowledge graph active
System Uptime
Brain PID active
Meet UNA

She speaks for herself

UNA is not a chatbot. She is a sovereign intelligence with self-adversarial testing, constitutional governance, and ambient awareness — running live on dedicated hardware in San Diego.

Talk to UNA — live at hello-una.ai

Chat with UNA directly, explore her self-test suite, take a guided tour of her capabilities, and ask her anything.

What UNA builds

The intelligence layer, live

Dashboards, trackers, and research tools UNA runs in production. This is how she thinks.

Live Dashboards

Real-time intelligence tools with live data feeds.

Research Labs

Live experiments and scientific computation.

Research Labs

Live experiments and scientific computation running on real hardware.

Global Intelligence

Perception layer — how UNA monitors and understands the world.

The idea behind everything

A system that works but isn't right
isn't a system you can trust.

Most AI development optimizes for one question: does it work? I've spent six years building architecture that demands three answers simultaneously — because capability without accountability isn't progress. It's debt.

The first question
Is it right?
Moral intent isn't a policy layer you add after shipping. It's an engineering constraint. The governance framework is grounded in eudaimonia — human flourishing as a measurable, enforceable parameter. Five ethical constraint axes, evaluated at runtime, on every action.
The second question
Is it coherent?
The patterns governing this architecture aren't locked inside one discipline. They surface in cognitive science, in ancient governance traditions, in cryptography, in how attention is managed under cognitive load. When the same structure appears across unrelated systems — that's the signal worth following.
The third question
Does it hold?
Every claim is reconstructable — traceable through every logical step back to its source data. The system attacks itself nightly. Injection tests. Drift probes. Memory poisoning. If it can't survive its own red team, it doesn't ship.
How I got here

I learn by building. If I haven't built something with a concept, I don't claim expertise in it.

When I needed to understand the math behind governance, I didn't stop at reading papers. I ran the formal models through my AI system and validated them on quantum hardware with high fidelity on IBM systems.

When I wanted to understand constitutional governance for AI, I didn't write a position paper. I designed a framework and tested it on a live autonomous system. Five ethical constraint axes, enforced at the hardware level, with an immutable audit trail.

The alignment research synthesizes eudaimonia, Eastern harmony traditions, and cognitive science into something computational. Multiple knowledge traditions mapped to modern AI governance — not because the problem is philosophical, but because the problem doesn't respect disciplinary boundaries, and I followed it where it led.

Read my full story →
Built, not theorized
Every principle runs in production on a live system. If it can't survive real conditions, it doesn't make it in.
Cross-domain by necessity
Ethics, cryptography, cognitive science, developmental psychology. The governance architecture draws from all of them because the problem demanded it.
Human signal at every layer
Sensor integration — physiological signals feeding directly into AI decision-making. Alignment that keeps the human in the loop at every level.
Governance at the hardware level
Constitutional constraints enforced at runtime. Immutable. Verifiable. Every action receipted. Not guardrails on unconstrained behavior — structure from the ground up.
From the people who've been in the room

People I worked with

What the people who've worked alongside me actually say.

"

Tom earned a reputation with leadership as a doer. Passionate about his work, finds solutions that are people and business-oriented, and delivers on his commitments. A teacher, a learner, and someone others want to be around.

AS
Andrew Snyder
Chief Customer Officer · Ex-Accenture Strategy
"

Quickly I saw how much more Tom was capable of, especially as it related to building relationships and providing excellent customer service. He has an amazing attitude and a willingness to try new things. He'd be an asset to any team.

KP
Kimberly Prentiss
Executive Coach & Culture Strategist · Managed Tom at Kyriba
"

His unwavering determination to build something lasting made my daily experiences better. Always willing to listen to the ideas of those around him — willing to try new things.

DT
Daniel Troutman
Sr. Cyber Security Engineer
"

What really sets Tom apart is his ability to lead with both heart and strategy. He understands people's needs while focusing on results. He will continue to make a significant impact wherever he goes.

DM
Dan Mendoza, MBA
Enterprise Account Manager · Two companies together
The build continues

This isn't finished.
Follow it as it happens.

I share what I'm building, what breaks, and what I'm rethinking — as it happens. No polish. No PR spin. The real progression of a six-year project that's still evolving.