https://datalaria.com/en/tags/software-quality/ Recent content in Software Quality on Datalaria Hugo -- 0.148.2 en-US Tue, 06 Oct 2026 00:00:00 +0000 https://datalaria.com/en/posts/evaluating_ai_agents/ Tue, 06 Oct 2026 00:00:00 +0000 https://datalaria.com/en/posts/evaluating_ai_agents/ Why traditional unit tests and metrics like BLEU or ROUGE collapse when applied to autonomous AI agents. We engineer a 4-layer evaluation framework: tool trajectories, environment state assertions, calibrated LLM-as-a-Judge, and CI/CD red-teaming.