Full talk · edited transcript

Forecasting at Databricks: One Framework Behind Consumption and Infra Cost

Ginger Holt, Alex Wang18 Jun 2026

Ginger Holt and Alex Wang, data scientists at Databricks. Databricks.

Both speakers are data scientists at Databricks and share a forecasting framework they built and use every day. The plan: first describe the framework at a high level (Ginger), then walk through the operational details of applying it at scale to infrastructure costs for engineering teams (Alex).

Ginger Holt has more than 25 years of forecasting experience in energy, finance, retail and technology; forecasting is often a first use case for data scientists working with business partners, and it never gets boring because things are always changing. Alex Wang has worked in the industry for more than 10 years, previously at Amazon and Lululemon, applying forecasting to demand and capacity planning.

Opening quote: "All models are wrong, but some are useful" (George E. P. Box). What Box did not say: even useful models get rejected by the finance team if they cannot defend the numbers in a CFO review.

Continue reading — it's free

Google or a one-time email code. No password.