In my case it was three main stages, firstly an assignment based assessment - 3 hours to complete a simple modelling and evaluation pipeline, including data preprocessing, feature selection, model choice, evaluation.
Secondly, a live interview with 2 members of Model Validation department with questions about different regression types, sql and simple data science.
Lastly, an interview with two team-leads focused mostly on an individual fit withing a given cluster in the team.
