Blog

I just finished the "Operationalize Machine Learning Models" module of the Azure AI 300 path. Here's the real take — scikit-learn, MLflow beyond logging, hyperparameter tuning as craft, and one lab that was confidently wrong.

I named my robot Sancho. Not by accident. Sancho Panza is the ideal AI — practical, grounded, loyal, useful. The Don is what we keep building instead.

Not a tutorial. A real account of how I use Claude Code day-to-day — what actually works, what I've stopped doing, and the three practices that changed how I build software.

A physical robot, built at home, running in my living room. Voice, memory, smart home control, animated eyes — all wired through the Claude SDK. Here's how it actually came together.