About Bill
Who I am
I am Bill Gernert, an infrastructure engineer in the Denver area. I spent fifteen years as a Windows administrator in higher education and county government, and I liked the Linux side more the whole time. In 2026 I stepped out of system administration on purpose to do platform engineering. I knew I could do the work. What I wanted was room to do it properly, so I built AutomationLab.
What excites me
The moment a platform starts paying for itself. The first server I provisioned by hand in the lab took an afternoon. Now a pull request does it, and I read about it in a Slack message afterwards. Certificates renew before I notice they are close. A restore check runs every week, whether or not I remember it exists, and tells me if the backups are lies. None of that is exotic. All of it is the reason I wanted the job.
Greenfield work. The lab was my first environment built from nothing, and I can see the seams. The roadmap and postmortem pages say what I would do differently. I want the next one, and the one after that, and I want each to be better than the last.
A team. I have been writing code for about eight years, and the best of those were the few I spent next to a good coder who was willing to mentor me. Most of the rest was being the one who fixes it. What I have not had is a group of engineers who own the platform together and argue about how it should work. That is the part I am looking forward to most.
Why I built the lab
I automated as much of my last job as the job allowed. I grew a Jenkins platform from 20 scripts to 135 Python and PowerShell jobs, the service desk could handle far more on its own, and after-hours work dropped. And the team was still a system administration team, measured on tickets closed and systems kept up. That is the right measure for that team. The change I wanted, where the pipeline is the product and the ticket is the exception, is a different mandate rather than a better script. It has to come from a team that owns the platform.
Some of what I wanted could not be built there at all. The systems were too far apart, and the change process was built for people rather than pipelines. So I built a place with nothing in the way, put my own name on the public side of it, and found out what happens when there is no one upstream.
What the lab taught me
A platform changes the pace of everything. When inventory, secrets, the pipeline and the monitoring read from the same source, one project stops waiting on another. I was carrying four or five major systems at once, and it worked, because each one could trust the records the others left behind. Server provisioning, certificate renewal, the restore check, the NOC and the Jenkins delivery platform all moved in the same stretch of weeks. I had never worked that way before, and I do not want to go back.
Efficiency compounds. Every job that ran on its own gave back the hours it used to cost, and those hours went into the next job. Triaging alerts, fixing a broken pipeline, sharpening a dashboard: all of it got cheap enough to stop competing with the projects. The more the platform finished, the more I wanted to build on it.
Blockers are everywhere, and they are rarely the code. They are the missing record, the credential nobody rotated, the approval that lives in one person's head. In the lab, the biggest blocker is me. I own every system, review my own work and hold every credential, which is a security flaw as much as a bottleneck. The delivery platform's review gates and the accepted-risk register exist to bound exactly that.
Security means something different when the public websites are yours. Every exposed endpoint and every default I forgot to change was mine to find, and I found several. The Security page is the record. The public status host now has no login path for anyone, and I am prouder of that sentence than of most of the features.
AI belongs where it makes the work more accurate, takes the repetition, and helps me think: brainstorming, research, a second set of eyes on a design. Repetition is also how people learn, so I am deliberate about what I hand off. Pushing it means finding where the guardrails need to be and what autonomy really costs. The Applied AI page shows where I landed, including the postmortem for the day I got it wrong.
I brought the Windows side into the lab on purpose. Active Directory, AD CS, Entra sign-in and the Windows template factory are here because hybrid shops still run them and I know them cold. But Kubernetes, Argo CD, cert-manager, Vault and the cloud-image provisioning path are the part I rebuilt most often and enjoyed most. Windows automation fights the tooling. Linux automation is the tooling.
The lab moves faster than this site. Pages distinguish live work from designed or exploring work, and live measurements are generated from the operating systems where that evidence exists. The site can still lag the lab, and that is a fair trade for the lab being ahead.
Where I came from
Before the lab I ran two Windows-heavy estates, one of about 1,000 VMs with more than 700 Windows servers and one of a few hundred servers, operated a 600 user Horizon environment, wrote a billing check that recovered about $10,000 a month in unbilled compute, and wrote the PowerShell service that watched 350 virtual desktops so we never had to buy the $30,000 a year product. I am a veteran. I served ten years in the Colorado Army National Guard, enlisted first and then through OCS to a commission, and finished as executive officer for a 90 soldier aviation company and then a 44 soldier space company, including moving 100 pieces of equipment across five states for a deployment. That is where the habit of writing the rollback before the change comes from. B.S. in Information Systems and M.S. in Finance, University of Colorado Denver.
Contact
My full work history is on LinkedIn, and that is where to reach me.