TabH2O Hands-On One API Call vs. the Pipeline You'd Build

Our first meetup introduced TabH2O, H2O.ai's tabular foundation model. This time, bring a laptop. You'll make live predictions on real Kaggle datasets with a single API call, turn on the new feature importance and anomaly detection capabilities, then build the equivalent LightGBM/XGBoost pipeline by hand — and see exactly what the foundation model absorbs for you: the type cleanup, the encoding decisions, the leakage traps, and the ~100 lines of code. We'll close the loop by handing the whole workflow to a coding agent with a single prompt.