The business case with a data set of 30k of observation with inbalanced data, with a long list of items to do, one of them being to run a "deep learning" model. When I tried to argue which data would make this model overfit and there it would not make sense, the "Data Scientist" replied, that is only for benchmark. They dont respect the candidate time, asking a lot of non-sense, and do not have knowledge to debate about the theory on how that would work or not