Free interactive tools
A production ML system you can click through
One real model, a predictor of which free trials will convert to paid, and everything it takes to keep its predictions flowing. Open the system that runs it, the repository it lives in, and the six months after launch when its world starts to move. Everything is clickable, and there is nothing to install.
Production ML
One real model, from training to the months after launch.
Anatomy of a Production ML System
A clickable map of the whole system, end to end: training, tracking, CI/CD, serving, monitoring. Open any component for what it does and why it exists, play a flow to watch data move through it, watch one break quietly, or step the build order to see it assemble.
Anatomy of a Production ML Project
The same system as it lives in git: the folder layout MLOps-mature teams converge on. Every file and folder is clickable for what it holds, why it exists, and where it shows up on the map.
Data Drift Simulator
Six months of the deployed model's life, played day by day. Nine short steps show what a drift check sees, when it fires, when it fires for no reason, and what it can never see. Then every setting is yours to try.
Walk the whole path, one part a day
MLOps for Data Scientists is a free five-part course: each part a practical guide paired with a working code repo, delivered to your inbox one part a day. Structure, serve, track, automate, monitor.