In a data science interview at a respected company I received a take-home with the framing “A Product Manager wants a dashboard with a massive amount of features (including interactivity, model prediction and GIS)” and received only 48 hours to do it. I thought there was some weird trick because that assignment had an unreasonable amount of scope in that timeframe. I found out the company uses a BI tool to streamline such tasks that is only available to enterprises and not consumers.

I eventually made the dashboard but it took 16 man hours; on the on-site, the interviewers implied they didn’t like it as it was not feature complete. (I called them out about the BI tool; they weren’t happy about it but admitted I was correct that it would be more efficient)

Now that I have had more experience as a data scientist, the real-world response to such a framing is to push back against the PM and write an implementation spec with a defined scope.