Metrics Driven Machine Learning Development at Salesforce Einstein
summary
Covers how Salesforce Einstein’s automated ML pipeline (Einstein Prediction Builder) uses monitoring and alerting to make a black‑box AutoML system more diagnosable. Describes tracking data and modeling metrics at each pipeline stage to detect issues, raise production alerts, and drive development of new features in a data‑driven way.
AI-written summary · about
chapters (6)
- 0:00Start
- 1:55Einstein Platform Machine Learning
- 5:23The Metrics Driven Approach to Model Development
- 25:37Experimentation with Label Leakage Detection
- 28:57Key Leamings from Modeling Experiments
- 31:40What Should We Track with Modeling Dashboards?
