Route optimization for field sales teams. Reps were burning gas driving random routes. I built the system that fixed it.
Sales performance tracking with gamification. People logged their numbers in Excel. Most didn't bother. I made it take 30 seconds.
Full-stack storefront with marketplace automation. Off-the-shelf covered 30% of what the business needed. I built the rest.
Real-time Voice AI orchestration platform. Sub-200ms conversational agents over streaming WebSockets.
I was 13 the first time I changed code to see what would happen. It wasn't programming — it was curiosity with a keyboard.
For years I learned in circles — new framework, small project, abandon, repeat. Collecting knowledge, not shipping anything.
Two years ago that changed. I started finishing things for real people with real problems. That's still the only reason I build.
If a user needs to read docs to figure out the next step, the software failed — not the user. Every click should feel obvious. Every outcome should be predictable. I don't care how elegant the architecture is if someone opens the app and gets confused. That's the whole job.
In the AI era, software that makes humans manually move data between tabs is already obsolete. The work worth doing is the work a machine can't verify — the judgment call, the relationship, the decision that needs a human in the loop. Everything else should run without you watching it.
Every project I've shipped started with something broken right in front of me — a process nobody questioned because they were too busy doing it wrong. The best software I've written didn't come from a product brief. It came from watching people work and thinking: this shouldn't be this hard.