ai-evaluation

Systematic evaluation (Evals) for LLMs and AI products. Design test cases, measure accuracy and quality, track regressions, benchmark models, and build continuous evaluation pipelines. Distinct from traditional software testing because outputs are probabilistic.