Cell 01Sovereign · isolated
- Own databaseisolated data
- Own encryption keysno shared secret
- Own AI endpointdedicated inference
No shared surface with any other cell.
Origin · the pre-interview
I’ve spent thirty years building the future before the market knew it wanted it.
Told the way I’d tell it on a first call, not a résumé — what I built, why security was never optional, and how the same instinct led straight to what I’m building now.
Act I · Ahead of the Curve
Ahead of the curve.
I never built for where the market was. I built for where it was going, and I got there early.
- A compact RFID/UWB tracking system — fifteen years before Bluetooth Low Energy made it a product. Sold that company before I was 25.
- The first unlimited wireless plan, eighteen months before anyone offered one commercially.
- The largest wireless ISP in the country: 175 square miles of T1 speeds to rural America while everyone else was on dial-up — a year before Google tried municipal wireless. Sold before I was 30.
Act II · Security by Requirement
Security by requirement.
At thirty I became EVP of a publicly traded company, Phazar Corp — and I stopped selling companies and started patenting systems.
- Hardware and software for the first large-scale secure mesh wireless system, and a $32M contract with the US Army.
- The first secure mesh optical relay for the Army: real-time visual confirmation, facial recognition at the edge — built to reduce the kill chain for embedded operators.
- Then the first end-to-end encrypted messaging built for ATAK — I drove the requirements and led the team; used across military and federal-civilian operations, through to its acquisition by AWS.
You build security differently when the people using the system can die if it fails. I learned that standing next to those operators, not in a conference room.
Act III · The Same Instinct, Now AI
The same instinct, now AI.
For the last five years I’ve been helping AI companies sell into the hardest customer there is — the DoD and federal-civilian world. And for four of those five years, I’ve been building my own LLM/AI answer. Its name is ZEUS.
I hit a frustrating wall — corporate guardrails that kept me from really meeting the operational requirements of the customer. That frustration led to a bias for action, and Zeus was born.
Because I originally built the LLM for me, I used my call sign — “ZEUS” — and it stuck.
That’s the story. Here’s what it built.
The proof · what the forging built
An engine, forged to be trusted.
Roughly three and a half years building one thing: an AI engine specialized for constructing real business software — the kind of system regulated industries could finally put their weight on.
Cell 02Sovereign · isolated
- Own databaseisolated data
- Own encryption keysno shared secret
- Own AI endpointdedicated inference
No shared surface with any other cell.
Cell 03Sovereign · isolated
- Own databaseisolated data
- Own encryption keysno shared secret
- Own AI endpointdedicated inference
No shared surface with any other cell.
Own database · own encryption keys · own dedicated AI endpoint — per customer. Sovereign-cell trust at multi-tenant cost.
Sovereign by design.
Every customer is a fully isolated cell — its own database, its own encryption keys, its own dedicated AI endpoint.
Trust the regulated can accept.
Insurance, healthcare, finance — the industries that couldn’t trust shared-tenant AI now have a model with no common surface to leak across. That boundary is what makes the answer “yes.”
Sovereign-cell trust at multi-tenant cost.
Single-tenant isolation at multi-tenant economics. Trust shouldn’t cost extra.
Proof of method.
One operator with AI leverage — and systems clients are still running long after sign-off. That’s what the thirty years were for.