Story
Twenty years of keeping systems running. AI is the newest one.
Mark Dalton has spent his career on the part of IT nobody notices until it breaks. It’s the part AI projects most often skip.
He started in product support and systems administration, and moved into virtualization when that was the big change in how companies ran their computers. At one company he cut physical servers by 75% and saved more than $1.3 million. Later he designed a system that ran more than 7,000 virtual desktops.
That era is where the name comes from. From 2015 he led vDesktop as managing partner and lead architect, building infrastructure for city governments. After that he ran infrastructure for other companies. In one role he took uptime from 89.7% to 99.9% across three data centers and two clouds. As a director, he moved a platform onto a new environment and saved $500,000 a year. In 2026 he certified as a VMware Cloud Foundation Architect and Administrator.
Where AI fits
He brought AI agents onto his own systems: running vDesktop on them, and building helloim.ai with them. He found what most people find. Getting AI to do the work is the easy half. Knowing it did the work right is the hard half, and it’s the same question he’s answered about servers for twenty years.
What he does now
He still designs and runs infrastructure, including VMware Cloud Foundation, and now helps organizations put AI into production on infrastructure they own. He also teaches people to do the same for their own work, one job at a time, with a check at the end of each.
From virtual desktops to virtual employees. The habit never changed: don’t assume it worked. Check.
How we build
The same checking we teach, on our own work, with the failures left in.
Case studies
The roles we never hired, and the weeks they got it wrong.
The method
How every agent’s work is checked before it counts.
The helloim.ai build record
What broke in our own app, what caught it, what it took to fix.
What we track
The benchmarks, models and research we build with, with sources.