Leesburg, Virginia · Washington, DC area
Endpoint, Identity and AI Systems Engineer
I've been the person people call when the computer won't do what it's supposed to since I was the Mac guy in high school. Right now that's at NPR, where I run our remote support platform and I'm leading the rollout of Intune to our phones and tablets.
I think in systems and processes. If I have to do the same thing twice, I'd rather build something that does it for me. More than 25 years in enterprise IT, and I still like figuring out why something broke.
This box will be an AI assistant you can ask about my work. It only knows what's in a file I wrote, and it won't guess past that. It runs on my own server first, uses a hosted model when it has to, and tells you which one answered.
It isn't live yet. Until it is, message me.
I took over NPR's BeyondTrust Remote Support a few months after I started, then moved it from the on-premises appliance to BeyondTrust Cloud. I planned it, set it up, tested it, did the cutover, and wrote the documentation. I still run it: who gets access, what gets logged, and how support sessions work.
NPR had no management on its phones and tablets. I ran the team that picked a platform. We tested three finalists head to head, and I pushed to judge them on features before price. Intune won because it can pull company data off a phone without touching anything personal, even when the phone isn't enrolled. Now I'm leading the rollout for both personal and company-owned devices, including how we remove company data when someone leaves. It isn't finished yet. On the Windows side, I'm in the Autopilot pilot group.
Vesper is the persona I run on Hermes Agent, an open-source AI agent from Nous Research. What I built is the second brain around her, which I started in January 2026, months before Hermes came along. There's a working vault for current projects and a separate archive of my history that the AI can read but not change. A guard blocks any write headed to the wrong place. There are rules for how she hands work to other AI tools and checks what comes back, and anything touching money, security, or deleting things waits for me.
Nothing counts as true until it's checked against a source. I designed it, and AI did most of the heavy lifting on the build. I also had Hermes taught to read more AI providers' "try again at" signals, so it switches back to the main model on time. That change was checked with 283 automated tests before I relied on it.
I started the move off our old shared drives, cross-trained with the systems team for three months, and built the PowerShell pieces for onboarding and migration. I wrote the guides for IT and for staff, ran the pilot, and handed it to the systems team. They finished it in spring 2026.
I ask the hard questions people usually don't want to answer, and I try to walk in with a fix already worked out, not just the problem.
I read before I touch anything. I'd rather check the vendor docs and the logs than guess. If patching something keeps failing, I rebuild it. Big changes get a test plan and a way to back out.
Before I trust anything important, I attack it four ways: a red team to break it, a blue team to check the defenses, a devil's advocate to argue with every claim, and a Gordon Ramsay pass for anything sloppy. When an AI tool guesses, I send it to the official docs.
If I have to fix the same thing twice, I automate it or write it down. Work smarter, not harder. A process has to have a point, though. If it doesn't make the work easier or better, it's just bureaucracy, and I have no patience for that.
Databases and SQL aren't my strong suit. I'll tell you that in the interview too.
Promoted to Specialist 3 in December 2025. Platform owner for BeyondTrust Remote Support, lead for the mobile Intune implementation, and escalation point for Windows, macOS, mobile, and identity issues. Active Directory, SCCM imaging, Iru (formerly Kandji), Okta. Trains new team members and supports executives and board members.
Earned CompTIA A+. Built Active Directory, DNS, DHCP, and Windows lab environments.
Tier 3 escalation in a secure biomedical research environment. Executive support, imaging to NIH security standards, surplus procedures and secure data destruction, security reporting. Peer coach for the team.
Phone, chat, and remote support for House offices. Knowledge base and install procedures.
Tier 3 escalation across Windows, macOS, and mobile.
Enterprise mobility management for iOS and Android. Wired, wireless, and VPN troubleshooting.
Contract-to-hire, converted to employee in November 2011.
Three roles in three years, each a step up. Multimedia and video production. Adapted Blackboard to run fully offline on remote computers using Apache.
At AOL: senior Mac and Windows support, product beta testing, and training new Mac consultants. Web hosting and domain support, on-site Mac repair, and customer care.
Ran an independent Macintosh consulting business (1993 to 1999). Medical specialist with flight medic training, Florida Army National Guard (1991 to 1993).
Only CompTIA A+ is an exam-based certification. The rest are completed training courses.
Florida State College at Jacksonville: completed the core classes of the Multimedia / TV and Video Production program (degree not awarded). University of Phoenix: business communications coursework.
I'm looking for a job where endpoint, MDM, identity, or systems work is the main job, not a side duty. AI enablement work that needs real guardrails interests me too. Remote, or hybrid in the DC area. Direct hire or contract-to-hire.