JEV TYPESAFE con SPRING AI - Tutorial Completo Fácil
Programando en JAVA walks through using Jev to judge an answer generated through Ollama. The Spring AI example combines relevance, quality and category decisions with a separate check that the answer is not empty.
Original by Programando en JAVAIntegrationsIntermediate14 min 17 secPublished
Generate the answer separately, then ask Jev whether it addresses the question; the walkthrough uses a 0.80 relevance cutoff.
Define ordered quality levels before scoring; this example uses five levels from zero to four and requires at least three.
Distinguish Jev category selection from the ordinary-code check for an empty answer; that check is not a fourth Jev question type.
Worth knowing
The presenter discusses library version 0.2.0 and uses Ollama to write the answer. Cutoffs are examples, not validated quality or safety guarantees. The later Java-category result is narrated as passing despite earlier naming Spring categories as acceptable. The implementation and speed claims were not independently tested. Spanish-language video; the English notes paraphrase Spanish captions.