I recently just got back a take home project for a start up, and the feedback I got was I don't deal with ambiguity well enough. I can't do anything actionable with this advice. Should I be asking more questions? Less? Better? Have a better understanding of what to do? What did you guys do to handle ambiguity better?
Should I be asking more questions?
Yep, especially given that this is an interview project. Take-home projects are a big enough time-suck as is; you don't want to spend 8 hours building out some sprawling project just to realize that 50% of the features weren't necessary or that you didn't use the libraries they wanted you to use.
If they don't answer your questions clearly, just propose an approach and ask them for a "yes/no". If they refuse to answer this or answer "No" without explaining what they want instead, this is a terrible company not worth interviewing for.
What did you guys do to handle ambiguity better?
I ask questions and break down the work, the latter of which I cover in-depth in this discussion.