The founder
Lee “BLT” Thomas
Founder · Sandy Hook, CT
For fifteen years before Zambarch existed, I put probability to work where being wrong costs money: first poker solvers, software that computes the game's best play via machine learning, the same family of computing as today's AI, then the daily fantasy models that produced a $1M win on DraftKings. The machines gave answers. My job was knowing which ones to trust. That discipline is the one I bring to every AI stack I read.
The skepticism is earned too. At Wise Woofs, the canine education company I founded and ran, we put the first public ChatGPT to work, and the early video AI for b-roll alongside it. Neither could stand next to what our people produced. So I treated AI the way I treat any unproven line in a model: set it aside and watch, returning only when the results hold up. They finally do, when the structure around them is right.
MIT Sloan’s “AI for Business Strategy” put formal structure under that judgment. The standard itself was set earlier, watching good people outperform the tools they were promised.
I was the AI skeptic in the room. I still am, until the math says otherwise.
The name
Zamb
From the Zamboni
Between periods, the Zamboni resurfaces the ice so the game runs fast and clean. That is the promise: a firm that moves on faster signals and smoother connections, with results that show it.
arch
From architecture
The structure underneath decides how well everything on top performs. We build the architecture that lets the AI you own run at its best: time saved and money made, by design.
Our name says what we do.
Start
Find out what your AI actually returns.
A few details, and I'll be in touch to set it up.
Or email directly: [email protected]