About

Portrait of Pharos Systems' founder

I get asked to build things fairly often. Only some of them actually get built.

A friend of mine designs commercial kitchens for a living, working with architects around San Diego. At a party at his house, as my family and I were getting ready to leave, he asked me: "Do you still write software?" He explained that the platform he depended on to spec products had been bought by a holding company that stopped answering support tickets — he was back to hunting through manufacturer PDFs and dealer catalogs by hand, trying to find something that actually fit a project's dimensions and budget. My honest answer in the moment was "let me think about it" — the kind of answer that usually means an idea isn't worth the time. But what stuck with me was what he was actually asking for: a catalog that was really a catalog of separate sources of truth. At Qualcomm, my team and I had helped migrate on-prem ServiceNow instances to the cloud, working with service teams across the company to get their day-to-day source-of-truth data into one place. I recognized the same problem here — automate the collection of data from multiple public sources of truth, normalize it, and present it to commercial kitchen designers. The real question was how to do it without incurring real hosting costs. Once I'd figured that out, I went back to my friend with an actual "I can do this." That became Pharos Kitchen Design.

Pharos didn't start as a personal project — it started as infrastructure. Once I was actually building PKD, I realized I'd need a small but real enterprise environment to support it: metrics, chat, a UAT environment, more than one DNS server with split-horizon capability, an authorization and authentication service, more than one virtualization host. I didn't know exactly how many services or hosts I'd end up running — I just knew I would, and that I'd need a way to track all of them. At Qualcomm, I'd used an internal implementation of RFC 2378 a team there had built, and I missed it the moment I needed the same thing for my own fleet. So I wrote an open-source version. I've talked to other former Qualcomm engineers since — more than one has told me the same thing: the one tool they missed most at their next company was that exact RFC 2378 implementation. Pharos exists because I needed it, and because I already knew I wasn't the only one.

Pharos Advanced Blocking started even more personally: managing my kid's internet access with Pi-hole and Unbound. When I needed a highly-available DNS setup for Pharos Systems itself, I landed on Technitium — multi-node HA, split-horizon DNS — and kept using it for my child's access too. The one thing missing was an easy way to edit the JSON blob Technitium's Advanced Blocking feature runs on. So I built a TUI for it. Same instinct as the other two, just aimed at myself first.

I've spent 15-plus years writing backend systems and leading the people who run them — most recently as a Senior Staff Software Developer at Qualcomm, leading development across 30-plus applications and services and mentoring teams of up to 14 people. Before that, a team of 7 at a digital-media company. None of that changes with the products — it's the same rigor whether I'm building for a friend, for myself, or for whoever finds this on GitHub.

Three products, three different starting points — but the same question decided whether any of them got built: is this one person's problem, or a lot of people's? If it's just mine, I keep it to myself. If it's not, I build it — whether that means pitching it to my friend's whole industry or putting it on GitHub for other engineers. Whichever one you're here for, that's the test it passed.