The Paris Machine Learning Study Meetup group held a mock interview. The purpose was to uncover the blind spots in our knowledge. We found some. I’ve posted a pdf of the list of questions. There are hundreds of questions cut-and-pasted into the document from various sources. The questions asked are checkmarked. The mock interview took two hours. Surprisingly, the most difficult questions turned out to be providing clear, concise definitions to basic questions such as “Define logistic regression?” or “Define overfitting”. The answers tended to be operational definitions and not proper definitions.
I think some people are having trouble reading the pdf because it’s composed of scanned images. I can rotate it to horizontal in any of 5 different pdf readers/devices (Android, Edge, iOS, Sumatra, etc.) but when it’s saved it reverts back to vertical. Since the pdf is a Dropbox share link, you might need the Dropbox app for access?
great idea!
if I can annoy you, I’d suggest rotating the .pdf so people can read on screen (I didn’t download, just tried to view in browser)
I had an interview (tech, but not data scientist) the other day. They asked to describe how I dealt with communicating to (agitated) non-tech people, who want answers and explanations (e.g. why things are taking so long, costing so much, not working as they should etc - all the things we imagine Jeremy has no problems with). Sure enough, followed up with “lets test it, explain what such_and_such is”