[D] How do you deal with unreasonable request from an employer with unrealistic expectations of ML?

Several months ago, I accepted a position to support a social science research project by training a ML model for them. The project involves using a dataset that the team (consisting of multiple interns, grad students, postdocs and professors) has compiled over several years and at an insane level of effort. However, the issue is that they failed to consult with anyone who actually knows ML beforehand. Their dataset is way too small (only about 200 rows) for what is …

dataset deal machinelearning multiple professors project research science social social science students support team them training