Career chapters

Experience

The Army made accountability concrete. Boeing taught me to see how local work fits into a larger engineered organization. At Church of the Highlands, broad systems responsibility grew into security engineering. At Waverly Advisors, that work now includes security-program leadership in a regulated financial-services environment. Independent software projects give me another way to test the architecture, authorization, and evidence questions that show up in operational work.

Engineering evidence

Selected engineering work

These projects make the design choices, limitations, and lessons behind the work inspectable. Full research publication remains on Aston Cyber where applicable.

Open-source demonstration

Security Triage Agent

A self-hosted demonstration of bounded AI-assisted security triage where model output remains advisory and deterministic controls govern evidence access, approval, and simulated action.

I built the project to test a narrow question: how can an AI model help investigate a security alert without receiving authority over security controls? Evidence tools are typed and allowlisted, model output is treated as untrusted advice, and deterministic policy binds every proposed action to explicit human approval.

The current release uses synthetic alerts and simulation-only actions. It is a demonstration of authority boundaries, auditability, and failure handling—not a production SOC platform or a claim of autonomous remediation.

View repository →

Graduate project

Secure Transaction Authorization Ledger

A graduate software project that binds approvals to exact transaction details and records authorization events in a tamper-evident hash-chain ledger.

The prototype explores how a firm can prove that the correct people approved the exact high-risk transaction under an enforced process. It canonicalizes transaction details, binds approvals to their SHA-256 digest, applies explainable risk rules, and preserves an authorization history through a hash-chain ledger.

It does not move money, connect to a bank, or replace operational judgment. The value is in making the authorization boundary inspectable and showing where software can strengthen evidence without pretending to eliminate human responsibility.

View repository →Read the research record →