David Green
Founder, RTP Applied Sciences
I’m an engineer who can run a sales call. Twenty years hands-on across healthcare, fintech, utilities and public sector: Veradigm, Envestnet, NetApp, Itron. I still write production code.
Much of that work was modernizing systems somebody else built, in insurance, utilities, healthcare and cloud infrastructure. The field notes are what those projects taught me.
Before any of that, psychology and economics at Northwestern. Then computer science with an AI concentration at DePaul, and an MBA at UNC. For a long time those looked like three unrelated interests.
Along the way I bought and ran a digital marketing agency, full P&L. Owning the number teaches you what a wrong decision costs.
RTP Applied Sciences is where the three degrees finally point the same direction. An old system is a record of thousands of decisions, most of them made by people who have since moved on, and the real process lives in workarounds nobody wrote down. Getting all of that back out before a rebuild takes all three.
When I’m not building, I ship small apps and speak at technical conferences.
Three disciplines, one question.
How do you change a system people depend on without breaking it? Psychology, economics and computer science each answer a third of it.
Psychology
How people actually work. The manual describes one process and the office runs another, built from years of small fixes nobody wrote down. Safety researchers call it the gap between work as imagined and work as done. And two experts reading the same rule will still disagree about what it means, which is why two reviewers decide separately.
Economics
What a wrong call costs, and to whom. A rewrite is one big bet on one number, and the planning fallacy says the number is wrong. Split into stages, each priced and each a place to stop, the same project becomes a string of small bets, and the arithmetic decides which one to place first.
Computer science and AI
How to read an old system fast. AI reads the code, drafts the rules it finds and writes the tests, which is where the speed comes from. A recorder shows which parts people actually use. A person decides every rule, because code can show what it does, not whether anyone still needs it.
The code is one version of the system. We rebuild all three.