Infrastructure automation and applied AI — for teams that ship.
I work with DevOps and product engineering teams: automating the toil nobody has had time to fix, and building AI systems that hold up in production. If AI isn't the right tool, I'll say so.
⬡ What I Build
Six kinds of work
Most engagements start with an audit. None of them start with a platform purchase.
Start here
AI Opportunity Audit
I learn your stack and your pipelines, then tell you where AI is worth applying — and where it isn't.
Build
Pipeline & Toil Automation
Log triage, alert noise, deploy checks, ticket routing — the repetitive work your team absorbs because nobody has had time to fix it.
Build
Agentic Systems
MCP servers, agent orchestration, RAG over your own corpus. Built to run in production and be operated by your team.
Build
Product Engineering Build
CI/CD and test strategy for teams whose velocity is capped by how long it takes to trust a change.
Build
Lab & Demo Platforms
Self-service environments wired into your CI/CD, so what gets demoed is the current build. I have built this at Fortune 500 scale.
Embedded
Team Ramp-Up
Months alongside your engineers, not a training deck. Someone in the repo who will tell you plainly when the answer isn't AI.
Built at Apstra · Scaled at Juniper · CCIE R/S #50117 (inactive) · Two decades in infrastructure
The work behind this
A handful of Ansible scripts became a Fortune 500's onboarding platform.
CloudLabs started as internal scripts that barely worked. I rebuilt it into a global, self-service platform — dozens of bare metal instances across multiple regions, wired into engineering's own CI/CD so what sales demoed was the current build, not a stale snapshot. It survived the Juniper acquisition and became the platform that ramped a Fortune 500 sales organization.
Four to five people ran it. That constraint is the whole point — I build platforms that a small team can actually operate, because I have never had the luxury of a large one. It is still running, five years on.
More work worth judging me on →Why Blue Ridge Systems
Scoped, not open-ended
You know what you are buying before we start. Every engagement has a defined boundary and a deliverable. If I think the scope is wrong, I will say so before we begin rather than bill through it.
You own everything
All code and deliverables are yours. No lock-in. No recurring fees for things you already paid for.
Right tool, not just AI
If AI isn't the right tool for your problem, I'll tell you that — and help you find what is. You will never get sold something you don't need.
Bring the problem, not the solution.
Thirty minutes, no pitch. If this isn't the right kind of work for me, I'll tell you that on the call rather than scope you into something that won't help.
Book a free 30-minute call