NNagappan S·Nov 4LLM-as-a-Judge: Using One Model to Grade Another (and Where It Fools You)LLM-as-a-Judge8 min read·nagspidey.hashnode.dev12S
AAdam Jaber·Oct 19LLM-as-a-Judge Bias: Why Your Eval Score Climbs While Your Answers Get WorsePosition, length, and self-preference bias in LLM judges, with the swap test, the pad test, and the calibration code to catch them11 min read·simplyexplained.hashnode.dev
MMuhammad Azlaan Zubair·Sep 25Jev Is Not a Paradigm Shift. It’s ust a Fast Classifier.Jev Doesn't Fix LLM Evaluation. It Just Makes It Faster16 min read·letinix.hashnode.dev22M
SSanath·Sep 24How BotGauge Is Building the Missing Layer for Trustworthy AI AgentsMost AI agent failures never show up in a test suite. Here's why red teaming closes that gap.9 min read·aiqa.hashnode.dev
MMuhammad Azlaan Zubair·Aug 22Your Agent Passed the Eval. Its Retrieval Calls Still Tripled.Task completion can stay flat while context compression shifts an agent's budget from execution to redundant retrieval.6 min read·letinix.hashnode.dev1
MMuhammad Azlaan Zubair·Aug 20Your AI Agent Eval Is a Production Security BoundaryA benchmark stops being neutral test infrastructure when the agent can execute code, cross trust boundaries, or affect real systems.9 min read·letinix.hashnode.dev32JK
RRedwan Rahman·Aug 10The Ruler Can Change Too: Navigating Judge Drift in Production AI EvaluationRedwan Rahman, Creator and Maintainer of NiriZan8 min read·nirizan.hashnode.dev12K
SShivam Bhardwaj·Aug 8My Reflection Loop Made Things Worse. My Evaluation Framework Showed Me Why.How building an evaluation framework before the agent revealed failures I never would have caught manually.11 min read·shivambhardwaj.hashnode.dev
AAlexander Codes·Jul 26Test Your Pydantic AI SQL Agent with Pydantic EvalsTest a Pydantic AI SQL agent with Pydantic Evals and evaldata. Check if queries have the same meaning, then compare their results when needed.7 min read·alexandercodes.hashnode.dev
VVamsi Krishna·Jul 18Day 56: Why evaluating the model is not enoughWhy evaluating agents means measuring the whole workflow users experience.3 min read·blog.vamsiannamreddy.com
