Adnan Masood, PhD.·Oct 20AI Metrics for Mere Mortals, Part 1: Why a 98%-Accurate Model Can Catch No FraudAccuracy, precision, recall, F1, ROC-AUC, MAE, and log loss, worked by hand: what each number means, how to calculate it, and when it…
bhupen·Sep 18The Faithfulness Metric: Catching the Hallucinations Your RAG System Doesn’t Know It’s Making !If you’ve built a RAG (Retrieval-Augmented Generation) system, you’ve probably had this moment: the answer reads perfectly. Confident tone…
Nitish Agarwal·Aug 5Stop Swapping Models. Build the Verification LayerTeams keep swapping models to fix flaky agents. The real bottleneck is that your outputs are probabilistic and your test suite assumes they…
InArtificial Intelligence in Plain EnglishbyAjay Verma·Jun 20Beyond the Assert Statement: Mastering the Art of LLM EvaluationIn traditional software engineering, testing is a binary world. If you input “2+2,” the output must be “4.” If it is not, the code is…
Alan Hsu·Jun 9Using LLM-as-a-Judge for AI Systems EvaluationThe Hardest Part of Building AI Systems Isn’t Generation. It’s Evaluation!
