I’m a developer who ships practical software for real ops workflows — AI agents, support tooling, and the glue that makes them trustworthy. I communicate clearly, own the work, and keep getting sharper.
I care about steady execution more than flashy claims. My focus is writing maintainable code, asking good questions, and creating momentum for the team.
I spent nearly four years in the United States Marine Corps as a Network Administrator, supporting mission-critical environments where reliability and accountability were non-negotiable.
After the Corps, I worked in technical support, then transitioned into software. That path gave me a practical mindset: solve the real issue, communicate clearly, and leave systems better than I found them.
"I bring discipline and consistency now, and I’m hungry to learn from stronger engineers every day."
Today I build AI systems and SaaS tools with TypeScript, React, and Cloudflare — especially products that sit where tickets, devices, and techs already work. I’m aiming for a role where I can contribute immediately while growing into larger ownership.
Experience highlights:
• USMC Network Administrator (2018–2022): operations, maintenance, and team leadership.
• Technical Support (~3 years): high-volume issue resolution and process improvements.
• Software Developer (present): AI agents, MSP tooling, and production SaaS on Cloudflare.
What I’m optimizing for: durable systems, clear ownership, and AI that earns its place in a workflow.
Two AI systems for MSP ops: Caisey puts a troubleshooting agent on the endpoint; Steno keeps ticket write-ups honest after the work is done.
Org-scoped fleet with manual deploy, auto-enroll, or client-specific installers. Runtime traffic is proxied for forward-only audit and conversation records. Permission defaults, Fast-mode controls, and multi-device prompts keep technicians in charge while the agent does the digging.
Confidence thresholds and reopen caps keep noise low. Per problem-type expectations replace one blunt checklist. Teachable in-ticket tags (##STENO_UPDATE##, ##STENO_EXPECT##, ignore tags) let techs shape behavior without touching the Worker. Dashboard + runtime on Cloudflare with D1-backed state.
Designed around a thin API layer and deterministic setup flows so support staff could handle onboarding without deep infrastructure knowledge.
Notes on AI in MSP ops — architecture, safety, and why summarizers never quite become technicians.
Architecture, information flow, independent risk evaluation, and the human-in-the-loop principles that keep technicians in charge.
Lots of MSPs are “trying out” AI. Very few are seeing it handle work — because context, access, and tools are still missing.