AI & Agent Systems
Systems that run on Monday, not just demo well.
- Agent architectures, harnesses & orchestration
- Governed autonomy & approval boundaries
- Retrieval & document pipelines
- Local & private inference
- Evaluation & honest assessment
About
The thread has always been systems.
Engineering led to infrastructure. Infrastructure made security unavoidable. Security sharpened the need for context. AI opened a new way to connect all of it.
The subjects change. The instinct—to understand the whole and make it work—does not.
DigitalPhysical
I build to understand.
My work sits where intelligent software meets the real systems it has to operate inside.
That means moving comfortably between the business problem and the architecture beneath it: models, data, networks, compute, security, operating processes, and the people responsible for the outcome.
Research matters because difficult claims deserve evidence. Building matters because architecture meets reality only when something has to run. Entrepreneurship matters because useful systems eventually have to create value for someone beyond the builder.
The discipline came the long way: Marine Corps data systems in deployed environments, then federal cloud operations, then consulting — places where the audit was real and downtime was not an option.
“The work is different at every scale. The discipline is the same: see clearly, respect the constraints, and make the parts cohere.”
I am equally drawn to digital and physical craft. A distributed agent system and a piece of furniture ask different technical questions, but both reward proportion, sequencing, material honesty, and an intolerance for hidden weakness.
2014–2020United States Marine CorpsData systems in deployed, classified environments
2019–2024U.S. Department of AgricultureFederal cloud operations on a 24-hour mission
2020-2023JRIT SolutionsPrivate Managed Service Provider for organizations throughout Long Island/NYC
2023–2025iTILITi Technology PartnersCloud Engineering and Solutions Architecture for Private Financancial Firms
NowM.S. Cybersecurity, USF & Several Startup InitiativesPost-Quantum Agent Harnesses, Next-generation EDRs, Private Intelligence
Capabilities
Six domains, one systems view.
The subjects the work falls under — each backed by the public record on its topic page, not by the list itself.
Cybersecurity
Security drawn from audit-grade and federal work.
- Security architecture & assessment
- Identity & access — Entra ID, PIM, Conditional Access
- SOC engineering — Sentinel + Defender
- SOC 2 Type 1 & 2 — led through audit
- DoD IA hardening — classified & unclassified networks
Infrastructure & Platforms
Cost-aware design that is still standing in five years.
- Azure & AWS architecture — landing zones, WAN, VPN
- Infrastructure as code — Terraform, Bicep, ARM
- Hybrid operations — 100% uptime, 24-hour missions
- Fleet management — Intune, 500+ devices hardened
- Deployable infrastructure — SATCOM, degraded comms
Engineering & Making
Digital and physical craft, same discipline.
- Full-stack systems engineering
- Hardware, prototyping & installs
- Field engineering — disconnected, degraded, deployed
- Woodworking that crosses into engineering
Technical Strategy & Leadership
The gap between executive intent and execution.
- Architecture direction & build-vs-buy
- Technical programs — $2M and $250K refreshes, end to end
- Led a 10-admin section and a 29-Marine platoon
- Operating models & governance
- Executive communication
Applied Research
Difficult claims deserve evidence.
- Experiment design & benchmarks
- Edge AI for cybersecurity — M.S. thesis
- Post-quantum cryptography · SCADA / ICS security
- Human-AI teaming for critical infrastructure
- Writing it up honestly
Operating principles
How I approach hard problems.
A compact set of ideas that stays useful across research, architecture, leadership, and hands-on implementation.
Evidence before assertion
Let the work, decisions, and outcomes carry the claim.
Systems before features
Understand the operating environment before optimizing one component.
Constraints are design inputs
Security, infrastructure, people, and failure modes belong at the beginning.
Build to learn
Turn an abstract question into something observable, testable, and useful.
The public record
This site is designed to show the work as it evolves.
Ideas begin as questions, become research or writing, turn into experiments and portfolio work, and sometimes develop a commercial trajectory. The links between them are part of the story.