Now

Updated November 2025

Current Focus

Building an AI assistant for executives — not for productivity theatre, but for real decision quality.

The challenge: make AI reliable enough that high-stakes users can trust it. That means solving the problems most AI products ignore — accuracy over speed, transparency over automation, systems that improve over time.

What I'm Learning

Building trust with AI — Not through marketing claims, but through architecture. What makes an AI system reliable enough for executives to depend on?

The gap between demos and systems — Demos impress; systems endure. Most AI products are 80 % demo, 20 % system. I'm focused on inverting that.

Why reliability is the moat — In AI products for high-stakes work, the competitive advantage isn't features — it's whether users trust it enough to bet their judgment on it.

What I'm Writing About

Lessons from this work — what I'm learning about reliability, decision systems, and building software that endure. Practical insights, not theories.