Closing a datacenter is a cost decision before it is a technical one. Ours was a colocation site, space we rented in someone else’s building, and it had changed hands several times over the years. Each new owner brought a new pricing structure, and the cost of staying there was climbing fast enough that keeping our hardware in that room had stopped making sense.

Most AI training is aimed at end users. The gap I kept running into was different. Executives and managers were being pitched AI tools every week, free trials were turning into production deployments without anyone noticing, and “I used Copilot” was being treated as a complete answer. The people responsible for the work coming out of their teams did not have a framework for asking the right questions. So I built one.

Most of my infrastructure work is about building things. This one was about refusing to let a vendor rebuild my cost structure for me. Two licensing fights landed in the same stretch, VMware under Broadcom and Veeam’s push off socket-based licensing, and both came down to the same discipline: know where the vendor is trying to take you, and know your own entitlements well enough to treat the contract as leverage instead of a formality.

Over about five years I ran multiple rounds of VMware host replacement and vSphere upgrades across our multi-site environment, and I did all of it during normal business hours with zero impact to the workloads running on top. Each round covered roughly 10 to 25 hosts at a time, full hardware refreshes, major-version ESXi and vCenter upgrades, and SAN refreshes running in parallel.

The Windows workstation imaging process was a bottleneck for our IT operations. It was manual, inconsistent, and disruptive to users. I saw an opportunity to change that. By leveraging Configuration Manager, automation, and the User State Migration Tool (USMT), I introduced a Light Touch imaging model that transformed the process into something efficient, scalable, and user-friendly.

Most internship programs start when a company has budget to spare and open roles waiting to be filled. I started mine during a downsizing, with a shrinking team and no management title, while most of the support work for four of our six manufacturing sites was landing on me.

Bringing 1Gbps fiber to four manufacturing plants reads like a procurement line item until you are the one signing the acceptance form. The plants needed the bandwidth to keep up with growing demand, the contracts specified 1Gbps, and Verizon was ready to call each circuit done. I made a validated throughput test, not a lit circuit, the condition of my signature.

The firewall project really started the morning of the ransomware incident, when one of the paths the threat actor used to move around the network turned out to be a legacy VPN appliance that nobody had been watching closely. Replacing the perimeter was already on the roadmap. After the incident it became the roadmap.

When a ransomware attack hit our organization, it was a wake-up call. It exposed vulnerabilities in our endpoint security and made it clear that we needed a stronger defense. Leading the charge, I deployed CrowdStrike Falcon EDR to transform our cybersecurity posture and ensure we were prepared for the future.

On a Saturday morning in September 2023, a Play ransomware variant encrypted virtually every Windows system across the company in about two hours. More than 500 endpoints and over 125 virtual servers were down before 8 a.m. I was the Service Desk Supervisor on paper. Within the first hour I was directing the technical response, and I ran it through recovery and the hardening that followed. We paid no ransom, lost no critical production data, and had roughly 80% of operations back within 24 hours. The work during and after the incident is what moved me into an infrastructure leadership role the following year.

Workstation refreshes are the least glamorous program IT runs and one of the most important. Done well, nobody notices. Done badly, every helpdesk metric gets worse for a year. Over more than a decade at the same manufacturer I ran several full refresh cycles across office laptops, plant-floor desktops, and quality-lab workstations, and the pattern that kept them quiet was the same every time.

The backup redesign was finished about a year before the ransomware incident that tested it. At the time it looked like a sensible refresh tied to a Veeam version upgrade. In hindsight it was the single decision that made recovery possible, and the difference between a hard week and a company-ending loss.

Migrating to a cloud-based helpdesk wasn’t just about upgrading software. It was about transforming how we delivered IT support. As Service Desk Supervisor, I led the transition from BMC TrackIT to Freshservice, introducing modern functionality, streamlined workflows, and a better user experience.