Abhinandan Kumar·6d agoBeyond Gini: Building an Effective Application PD Validation FrameworkFinancial institutions invest significant effort in developing Probability of Default (PD) models, but model development is only half the…
Ankitkumarjhajharia·Sep 30Why DHARA Uses Multiple Hydraulic Models Instead of a Single SolverA common approach to hydraulic modelling is to select one solver, prepare the input data, run the simulation and analyse the output.
Alan Scott Encinas·Aug 10Kaggle Leaderboard Shakeup: When a Metric Stops RankingEvery improvement I made in the back half of this competition was real, in the sense that the number went down. I moved from 7.623 to 7.618…
ait-lahssen Ismail·Aug 9When Should You Retrain a ML Model?Monitoring tells you something changed. Retraining decides whether your model deserves to keep making decisions.
Rach·Jul 21Validating Financial Encoder NLP Models under SR 26–2: When Accuracy Is Not EnoughIntroduction
Erika Gomes-Gonçalves·Jul 20Why powerful ML is deceptively easy — Part 2 | Towards Data ScienceThe next leakage problem is not only temporal. It is spatial, structural, and coverage-related.
Alan Scott Encinas·Jul 17The Validation Score That Lied to Me for a MonthMy local validation and the leaderboard were measuring two different statistics, so my home score kept lying to me. How I caught it, and the
AnyLogic·Jul 146 Common simulation design mistakes and how to avoid themThe AnyLogic Technical Support Team works with users’ simulation models every single day. We see everything — from well-structured models…
Alan Scott Encinas·Jul 8I looked unbeatable in practice and lost the real fightA change that gained on my home bench lost three times as much on the hidden leaderboard, and my own tooling couldn’t reproduce the failure
