DBTB INT Parag Namjoshi
summary
Workday frames ten years of transaction logs as product-differentiating data and shows a practical shift from commoditized systems to using those logs for models The core claim is that mistakes in validation methodology wreck predictive work and derail production deployments The talk lists concrete fixes for predictive classification pipelines, emphasizes training gaps in probability and statistics, and points to continuous practice and integration into engineering workflows as the path to reliable models
AI-written summary · about
related talks
23mDelivering Modern Data Stacks With dbt to Operationalise Analytics
01h 42mFall in Love with Data Transformation: DBT as Your Weapon
34mData Science beasts (failures) and where to find them
24mMake data central feature
44mPreventing, Diagnosing & Curing Bad Data
39mPanel Discussion
18mData quality: the key to long term happiness
28mDataEngBytes 2023 - PER-T1-05 - Ellie Gholami