About EthicVault

Advice flags. Infrastructure refuses.

EthicVault is not an AI policy tool. It is execution-layer infrastructure — a deterministic barrier that stops non-compliant AI actions before they reach production systems, with every verdict anchored to an append-only, hash-chained ledger.

The problem

AI acts. Governance watches.

Enterprise AI systems execute trades, transmit data, and modify records autonomously — with no deterministic enforcement between intent and action. Existing tools are advisory: they flag, score, and report. They do not stop. And because they run as software beside the model, the same input can produce a different outcome depending on the floating-point backend it lands on.

Our position

Permission is architecture.

When deployed with complete mediation, every in-scope consequential action must cross the governed authorization boundary — below the model, before the runtime. Verdicts are computed in fixed-point arithmetic against frozen thresholds, bound to cryptographic proof, and enforced without discretion. In our silicon line, the same verdict physically opens or interrupts the actuation path itself.

02The Team

Governance built by people who have run systems at scale.

Deterministic governance is not a research idea here. It is built by operators who spent their careers where delivery quality, risk, and accountability meet — in enterprise consulting, national-security research, and global supply chains.

Dr. Abdelhalim Rekab

Dr. Abdelhalim Rekab

President and CEO

Creator of the Resonance Engine (U.S. patent pending). A PhD in Sociology of Communication let him map the structural risk hidden inside neural networks; the answer required systems engineering, not behavioural filters.

LinkedIn
Sharaf Taha Alshaibi

Sharaf Taha Alshaibi

Exclusive Representative — Saudi Arabia & GCC

Saudi businessman based in Jeddah, with experience in business development and commercial relations across Saudi Arabia and the region. He is the sole and exclusive representative of EthicVault for the Kingdom of Saudi Arabia and the GCC, leading regional operations, strategic enterprise partnerships, and client engagements as the single official channel for deploying our deterministic AI governance across the Middle East.

Cris Peterson

Cris Peterson

Executive Vice President

More than 38 years directing sales, global operations, and complex customer projects and delivery strengthens the company's ability to build and scale reliable systems. His career has centered on creating disciplined, accountable processes that reduce variability, control costs, and consistently deliver large-scale programs across international markets.

LinkedIn
Nargesse Peterson

Nargesse Peterson

Chief of Staff

Built and debugged the integration layer behind multi-million-dollar logistics for BASF and Dow Chemical. Works across EDI, Kafka and cloud middleware; fully bilingual in English and French.

LinkedIn
Amine Rekab

Amine Rekab

Manager, AI Compliance Architecture

Runs deterministic execution control: the Friction Gate that intercepts anomalous agent behaviour before it reaches production — with a cryptographic audit trail behind every blocked action.

LinkedIn

Advisors & Counsel

Dr. Dwayne Rosenburgh

Dr. Dwayne Rosenburgh

Senior Advisor — Cybersecurity & Silicon Assurance

33 years leading RF, wireless and cybersecurity research for the U.S. government, where he founded and led a 20-person research division. Former Research Scientist at MIT Lincoln Laboratory and senior program leader on a ~$300M U.S. Cyber Command program.

LinkedIn
Edward P. Dutkiewicz

Edward P. Dutkiewicz

Patent & Corporate Counsel

Counsel to EthicVault on its patent portfolio and corporate matters, including the pending U.S. applications covering the governance method and its hardware embodiments.

03Intellectual Property

Filed first. Built to the filing.

EthicVault's architecture is the subject of pending U.S. patent applications covering both the software governance method and its hardware embodiments.

1.1

The governance method

A computer-implemented method for governing outputs of an AI language model: layered separation of computation from authority, fixed-point state, ordered validation, and proof-bound execution.

1.2

The hardware continuation-in-part

The same architecture enforced in silicon — governance SoC, hardened IP macro, and a hardware execution restrictor interposed in series with the actuation path of medical, automotive, and payment devices.

1.3

Reduction to practice

The fixed-point gate has been expressed in synthesizable HDL, verified bit-exact against a golden model, and carried through timing closure on FPGAs from two different vendors with identical dispositions.

Ready to see it in production?

We are onboarding a select group of regulated institutions. Request a private technical briefing with our engineering team.

Request enterprise access