Shady recruiting practices. Started out as a typical HR phone screen, however things went off the rails when the recruiter started asking me details of my medical history (I took a medical leave for about a year and was reentering the job market). I informed him that 1) if it was still a problem I wouldn't be applying to the role 2) it was none of his business anyways and that 3) this conversation was over and I was no longer interested in a position with a company that would ask for such information. It is illegal for employers to ask about medical conditions and it is also illegal for employers to discriminate based on medical conditions or disabilities. Avoid this employer.

The pay offered by this position is extremely low, only starting from 55K to 75K. I don't think they are hiring data scientists. HM who interviewed me only uses R in his work, but he will challenge you with some packages used in Python which he doesn't quite understand. While I was talking about my project, I told him only binary classification can be balanced its class weight by XGboost via scale_pos_weight. I have to write self-defined function to balance Multiclassification problems. But he insisted there is a tuning method in XGBoost to tune. He also asked me why I used MSE instead of using MAE, I told him it is easier to calculate R square, but he thinks MAE is more robust to outliers...... That dataset is a time series dataset with 200 data points, how many outliers you can expect to have in such small dataset?