Skip to content
← Back to writing

March 9, 2026

The real timeline

I built like a maniac for six weeks. Then I stopped for five months. Then I came back and shipped everything in three weeks.

July 31, 2025

Proof the work was real: the HomelabARR repo public on GitHub — README, MIT license, tagged releases, a project out in the world with its own domain.

I got laid off from Neptune Technology Group. I was a Software Project Manager leading 12 Scrum teams on their biggest platform. After nearly a decade in project management, I was suddenly looking for work.

I had some savings, a daughter to support, and my parents were generous enough to let us move in. That was the safety net, and there was no plan B.

August 2025: the sprint

Two weeks after losing my job, I started building. I already had a homelab running 34 Docker containers, and a CLI tool I had been tinkering with for managing them. On August 14th I wrote the first real documentation page. By August 19th I had written 18 Confluence pages in a single day, and 32 more by August 24th.

I rewrote the entire backend from JavaScript to Go in three days because I decided the performance was not good enough. I had never written Go before. I did it anyway because Claude could help me learn the syntax while I focused on the architecture. The PM skills turned out to be the real asset. I knew how to scope work, break it into tickets, and ship incrementally, and the language was just syntax I could pick up as I went.

I also started NeuroHelper the same week. An iOS wellness app. Breathing exercises, calming tools, AI chat. It started as a web app, then Expo, then I decided to rewrite the whole thing in Swift because I wanted it to feel native.

By September 10th I had completed a full modular architecture refactor. 56 story points across 8 stories in two days. 121 functions preserved with zero breaking changes. 135 Confluence documentation pages, and 238 total pages by mid-September.

I was building like my life depended on it because it did.

October 2025 through February 2026: the gap

Then I stopped.

I do not have a clean narrative for this part. My git history goes almost completely silent from mid-September 2025 through February 2026, five months of almost nothing.

The context matters. In the year before I got laid off, my divorce finalized. My daughter was struggling. The job ended and I was suddenly looking for work with a lot happening at home.

So when I say I stopped building for five months, it was not because I lost interest. I was a recently divorced single dad who had just lost his job, living in his parents' guest room, watching his savings drain, and trying to hold it together for a kid who needed her dad more than she needed him to ship a NAS operating system.

And those five months were not empty. I spent them resetting and getting my footing back so I could think clearly again. That reset was the prerequisite for everything that came after.

I am not going to romanticize this period or skip over it. It happened. Five months of not shipping code, and five months of doing the harder work, and that is part of the story.

March 2026: the comeback

I do not know exactly what flipped the switch. Maybe the weather changing, or my daughter starting a new routine, or just enough time passing that I was ready again. But in early March I sat down and started building again.

And then I did not stop.

In the days since I came back I have been shipping again. I launched a product site for HomelabARR with tiered pricing. I released the Community Edition as free open source with 137 app templates. I submitted NeuroHelper to the App Store. I started building a managed hosting platform and a lineup of AI agent template products. I am back to the work I love and the plan is to keep going.

I also have an AI assistant running 24/7 as an operator that monitors my infrastructure, posts to social media, and handles the operational overhead of running a software company solo.

The math

People ask how I shipped so much so fast. The honest answer is that I did not. I shipped in intense bursts separated by five months of nothing. The six-week sprint in August and September built the foundations. The March sprint is building everything on top of them.

Total actual build time so far: several weeks across eight months. I did not grind for seven straight months. Most of that time went to healing and to raising my daughter while I got back to a place where I could build at all.

I am not ashamed of the gap. The gap is part of the story, and it is what makes the comeback real instead of some LinkedIn fairy tale about hustle and discipline.

What I have now

As of today I have HomelabARR PE and CE, NeuroHelper in App Store review, a marketing site, a portfolio with a real blog, and an AI assistant that handles a growing share of the operational overhead, with more coming.

I am still living with my parents. I am still scared. A few sales have trickled in from strangers who found the product and paid for it. It is not enough to live on, but it is enough to know the mechanism works. I am still applying for jobs at the same time, because one sale a week does not pay the bills.

This is a work-in-progress story from someone who got knocked down, built like a maniac, crashed, and got back up. If it becomes a success story, I will write that post too. But I am not going to pretend I am there when I am not.

I am a Technical PM who taught himself to code with AI. I am a single dad. I build software from a server rack in my parents' house. That is who I am.

mjashley.com · HomelabARR