Built at the supply-house counter, not in a conference room.
I'm not a career field tech. I went to Universal Technical Institute, earned my 608 and NATE Core, and spent the last year on the counter at one of North America's largest HVAC/R distributors. What I kept seeing wasn't a lack of effort, it was the learning curve. So I'm building tools to bridge that knowledge gap and shorten it.
The long road in
Eleven years in the U.S. Army. Nine years in cybersecurity after that. Two careers where the cost of guessing is somebody's safety, and where you learn that the only thing that scales is a repeatable process.
Somewhere in there I decided I wanted to work with my hands again. I enrolled at Universal Technical Institute (UTI), finished the HVAC-R program, picked up my EPA 608 Universal, NATE Core, and OSHA 10. Took a counter job at one of North America's largest HVAC/R wholesale distributors, so I'd be around the trade every day while I learned it.
What I saw at the counter
Techs walked in all day with the same kinds of problems. "It's low on the suction side, must be undercharged." "Replaced the cap, still not cooling." "Owner wants the bill, but I don't actually know what failed."
Almost none of it was a knowledge problem. The techs knew the fundamentals. What they were missing was a process — an order of operations that didn't depend on remembering every PT chart, every superheat target, every signature of every fault, all at once, on a 100°F rooftop, while a customer waited inside.
The tools they reached for didn't help. They'd ask ChatGPT, Google's AI, or Perplexity — but a general-purpose engine with no real context about the system in front of them tends to confidently hallucinate a fault and let a tech change a $400 part on a hunch. None of them showed their work.
So I built tools that do
Every fault model exists because of a problem I heard a tech describe at the counter. Every UI decision is built for the truck — big, glove-friendly targets and high-contrast numbers you can read on a bright rooftop. Every diagnostic ranking shows the reasoning behind it, so the tech can disagree with the engine, leave the call right, and pick up the why for next time.
The diagnostic core is deterministic — pressures and temperatures in, ranked faults out, with the math and the indicators that produced them. It works offline because basements and rooftops don't have cell service. That conviction grew into two products: HVAC-R Diagnostic Pro, and SCSH, which pairs the same deterministic core with an AI mentor that explains the call like a 25-year veteran would.
Independent work
Determinari is independent work. It was built entirely on my own time and my own equipment, with no employer resources, systems, or proprietary data involved, and it has no affiliation with — or endorsement from — any distributor or employer. The trade taught me what to build; the product, the code, and the company are my own.