1st round was a Hackerrank Online Assesment, where you need to create a model on a dataset they give on a JupyterLab style IDE. You can choose what model you want to write, and you can import libraries etc. They grade on accuracy on a hidden test set, and code clarity, visualizations etc. Next round is final round.

Hackerrank type machine learning modelling question followed by two interviews, one focused on the modelling question done in the previous stage and the other on more general behavioural and technical questions