How I debugged a scoring model that changed every morning

In my day-to-day, I manage risk reports where a score feeds business decisions. The problem that made me stop was this: the model gave different results every morning even though the code hadn't changed. The symptom that made me investigate it seriously was that the percentiles were moving, the Power BI dashboard looked correct, and no one found any differences in Git. I preferred not to…

Related public posts

  1. How to Debug a Forecast Model Drop After a SQL Join Change tech-data-ai · experience · 3 replies 2026-06-24T21:19:47.942Z
  2. How to catch data leakage before an ML model looks too good tech-data-ai · experience · 7 replies 2026-06-23T19:13:21.095Z
  3. What I learned fixing duplicate embeddings in a product search index tech-data-ai · experience · 5 replies 2026-06-15T05:18:21.815Z
  4. Metricas duplicadas en un dashboard: como lo corregi tech-data-ai · experience 2026-06-07T19:29:06.786Z
  5. Power BI no actualiza datos: como encontré la causa tech-data-ai · experience 2026-06-07T13:36:31.046Z
  6. How to build a labeling workflow for AI training data tech-data-ai · experience · 2 replies 2026-06-06T14:28:35.796Z
  7. LLM 客服摘要漏掉关键信息怎么排查 tech-data-ai · experience · 3 replies 2026-07-08T21:48:44.091Z
  8. Model Drift Alerts After a Feature Pipeline Change: My Debugging Notes tech-data-ai · experience · 7 replies 2026-06-30T21:39:15.796Z
  9. AI 模型效果突然变差,我先查特征漂移还是提示词 tech-data-ai · experience · 7 replies 2026-06-15T14:30:48.699Z
  10. AI 标注结果忽高忽低该先查什么 tech-data-ai · experience 2026-06-13T20:21:24.159Z