production paths, after-hours labs.
ML systems and data pipelines by day; agents and tiny autonomous things by night. A few of the boundaries I've drawn well — and the stacks behind them.
currently building
The one I'm shipping right now — a practical loop for learning with an agent beside you.
A local tutor experiment for learning with agentic loops, tight feedback, and deliberate practice. Small enough to understand, useful enough to keep around.
the back catalogue
Agents, simulations, and the tooling behind the content. Most of it open on GitHub.
pi-tutor
A local tutor experiment for learning with agentic loops, tight feedback, and deliberate practice.
pi-troupe
TinyTroupe-style persona simulations for focus groups, interviews, validation, and market-research hypotheses.
pi-overwatch
Watchful automation around the Pi ecosystem — observing, reporting, and keeping small systems honest.