They Said It Was Impossible.
I Had 17 Days.
The story of Verora AI didn't start in a computer science lab. It started at a kitchen table on February 17, 2026 — with a father of three (and one on the way) who had never opened a terminal, asking a chatbot: “How do I create an actual software app with the files you sent me?” Twelve minutes into his first conversation with AI. Ever.
The Ivy League Dropout
I dropped out of a Master's at Columbia University — Enterprise Risk Management — to build this. Not because the degree wasn't valuable. Because the problem was more urgent. ERM taught me one thing above everything: you don't build features. You build risk-mitigation engines. So when we say Verora stores zero patient data at rest — encrypted Redis cache wiped at 11 PM every night, no PHI in the database, ever — that's not marketing. That's risk architecture. Assume breach, minimize blast radius, never store what you don't need. Four days into the build, before I even knew what a git commit was, I was already asking about BAA requirements and HIPAA audit logs. The instinct came from the training.
The Army Grit
I'm 33. A U.S. Army Veteran. I learned to operate in language barriers and cultural chaos — the kind of “miserable but fun” that only makes sense when you've lived it. That same grit is what carried the 17-day sprint. Day 0: I asked “the native terminal in my computer or the visual studio?” and npm rejected my project name for having capital letters. I named the folder pulpmain by accident because I couldn't figure out the naming restrictions. It's still called that. By Day 17: 315 commits, a production autonomous engine processing real patients for real dental practices, and a parallel overnight cron that scales to 400 offices. No bootcamp. No CS degree. Just obsession, a supportive wife, and a lot of late nights after the kids were asleep.
The Aha Moment (The Brother's Gripe)
The vision didn't come from a whiteboard in a WeWork. It came from gaming with my brothers — who work in dental tech. Hearing them say, over and over, that no one could truly verify insurance. That the whole industry was held together with fax machines, offshore call centers, and hope. That front desk staff spend two hours every morning on the phone with payers just to confirm what should already be known. The legacy players had decided 100% autonomous verification was impossible — that you'd always need a human in the loop. I decided they were wrong. Not because I knew more about insurance than they did. Because I came from outside and didn't know what was “supposed” to be impossible.
The Why
I've got three kids and one on the way. My wife didn't sign up for “husband disappears into a code cave for 17 days,” but she showed up anyway. Every feature in this engine was built after the kids went to sleep. The overnight cron that verifies every patient before 5 AM? I built that between midnight and 3 AM. The dependent verification system that catches the 30-50% of patients that every other tool misses? That was a Saturday. Building Verora didn't just change our product — it changed our life. The kind of focus that only makes sense when you know what you're building for: something that actually works for real practices, real patients, real insurance chaos. Not a demo. Not a pitch deck. A production system that's already running.
To the Dreamer Who Doesn't Know Python
You don't need to know Python. You don't need a CS degree. You don't need to know what a terminal is — I didn't. On Day 0 I was asking whether to use “the native terminal or the visual studio.” On Day 17 I was architecting parallel distributed systems. The gap between those two sentences is not talent. It's not money. It's not connections. It's obsession. Showing up at the kitchen table every single night after the kids are asleep and refusing to stop until the thing works.
If you have the obsession, the tools are already in your hands. The folder is still called pulpmain. Start ugly. Ship anyway.
Ready to see what obsession built?