Uploaded October 2026 | Updated November 2026, 11 hours ago
Watch Gabe Weiss go from raw Lakehouse discovery to a trained XGBoost model, a staged marketing campaign, and a zero-token BI dashboard in four minutes using Gemini and Knowledge Catalog. Watch the full Keynote: https://goo.gle/47cXmir
- Live PySpark on Lightning Engine: Grounded in Knowledge Catalog, Gemini writes inline PySpark in a live Python notebook—pushing zero-copy joins to Lightning Engine to find 52,000+ bank customers who can save an average of $3,400.
- Explainable XGBoost on Managed Spark: Trains an XGBoost propensity model on 3 years of BigQuery campaign history (90% accuracy, 85% precision), surfaces the top predictive signals, and publishes the model back to Knowledge Catalog.
- Marketing Execution & Zero-Token BI: Invokes the cataloged model in natural language to stage a campaign in Google Marketing Platform, then builds a live county dashboard that refreshes via Knowledge Catalog query templates using zero LLM tokens.
Watch Gabe Weiss go from raw Lakehouse discovery to a trained XGBoost model, a staged marketing campaign, and a zero-token BI dashboard in four minutes using Gemini and Knowledge Catalog. Watch the full Keynote: https://goo.gle/47cXmir
- Live PySpark on Lightning Engine: Grounded in Knowledge Catalog, Gemini writes inline PySpark in a live Python notebook—pushing zero-copy joins to Lightning Engine to find 52,000+ bank customers who can save an average of $3,400.
- Explainable XGBoost on Managed Spark: Trains an XGBoost propensity model on 3 years of BigQuery campaign history (90% accuracy, 85% precision), surfaces the top predictive signals, and publishes the model back to Knowledge Catalog.
- Marketing Execution & Zero-Token BI: Invokes the cataloged model in natural language to stage a campaign in Google Marketing Platform, then builds a live county dashboard that refreshes via Knowledge Catalog query templates using zero LLM tokens.










