Comment by ojas01

Interesting approach. I’ve seen a similar challenge when testing data pipelines where schema complexity makes simple fixtures unrealistic. Synthetic generation tied to schema definitions could help keep test environments closer to production structure.