Evals: how you actually know it got better
Feeling better isn't the same as measurably better. Offline eval sets, LLM-as-judge, human eval, and regression testing for prompts and RAG pipelines.
Blog
#eval
Posts on AI engineering, LLM systems, and software development.