Right now I'm the founder of ArticuLearn, where I'm moving LLMs beyond simple chat interfaces into reliable, context-aware agents that can facilitate real human growth. The hard problem isn't generating a response, it's persistence: long-term memory and personalized feedback that compound across sessions instead of resetting every time you open the app. I designed the tutor's memory architecture, built the voice pipeline, and shipped the product end to end.

I've shipped two others solo, each deliberately using a different tool. WarehouseWatcher is a real-time price tracker, live in both app stores, that delivers alerts without storing a single user record. Preference Scheduler is a workforce scheduling engine built on a constraint solver rather than a model, because a schedule that merely looks right is worse than useless when it quietly violates someone's required rest hours. Three products, three approaches, and each choice came down to whether the problem needed an answer that was plausible or one that was provably correct.

That judgment is what I think engineering leadership is now. In the age of LLMs the scarce skill isn't building with models, it's knowing which problems deserve one and building teams who can tell the difference between a demo and a system. I lead high-autonomy teams, I coach deliberately, I hire carefully, and I believe the best systems come from diverse groups willing to tackle the hardest "how" in service of the most impactful "why."

🧑🏾‍💻 Pronouns: he/him