The process went through a phone screening, a Skype-based high-level technical interview, and a Skype-based more detailed technical interview. We stopped the process at that point, although there was an initial plan for an on-site detailed technical interview.

The whole process has been fantastic. The HR manager was very diligent, the hiring managers were nice, open, and the best examples employees satisfied with their company.

I did not get an offer as another candidate was better for the position, but the whole experience has been positive. I appreciated the final announcement, detailed with clear reasons for their decision. I applause at such a rare and proper closure---an example many companies of this size should follow.

Given a Jupyter notebook with code to load a classifier and images, there is a bug on inference. Find it and explain.

Given some ML-related source file, explain it. What does it do? How does it fit a project?

How to deal with limited amount of training data?

How to deal with unbalanced data sets for classification tasks?