Governance and assurance for AI in regulated life sciences and healthcare. I write the specification and build the verification — and I publish what the audit of my own instrument found.
Agentic and open-weight model assurance. Evaluation design. EU AI Act, ISO/IEC 42001 and NIST AI RMF conformance. Albuquerque, New Mexico; remote, U.S.
Evaluation instrument v3 — designed and built to test whether AI delegates deliberating across an enforced context-isolation boundary produce valid synthesis: rubrics pre-committed, targets sealed, arms ablated, leakage reviewed, provenance exported.
EvidenceThe audit — an origin-blind adversarial review of that instrument; six defects verified in code; the register is public and versioned.
RegisterTension OS v2.1 — a decision-and-governance methodology with conformance cross-walks to the EU AI Act, ISO/IEC 42001 and NIST AI RMF; a working demonstrator and three public GPTs.
MethodologyWho this is for
Health systems and academic medical centers, pharma and biotech, payers, and the Responsible-AI and validation consultancies that serve them. Governance, assurance, audit and Responsible AI roles at manager to senior-manager level, remote; and contract engagements in evaluation design and conformance.
Where the judgment comes from
Cancer-biology research training and laboratory operations at MD Anderson, Albert Einstein College of Medicine and Identigene, with a co-authored paper in ACS Medicinal Chemistry Letters. Eleven years of healthcare-access venture work reviewed by hospital leadership, a payer and bioethics faculty. An Executive MBA. English and Spanish.
The work history carries the dates; the resume is the same facts in one page.