Right in March: Machine Learning in Production, Drift, Fairness and Explanation (Machine Learning series finale)

Deep End Era. Engineering with Sankalp · Oct 28, 2026 · 55m

A fictional building society's new loan model was right in March. By October nobody could say whether it still was, and one group of applicants was being quietly declined twice as often. Credit risk head Nadia Haddad and engineer Ben Ashworth find out why. The Machine Learning series finale covers drift and monitoring, feature stores, evaluation that matches the real decision, conflicting definitions of fairness and the COMPAS debate, SHAP and its limits, Amazon's hiring model and other famous failures, and what production really demands. Cross-references: ep116, ep176, ep198 and all parts…