Found the paper https://www.usenix.org/system/files/conference/hotcloud16/ho.... Puts forward a very nice overview of distributed container patterns. But maybe in articles that Google Scholar says cite this one, a name has already been found for this specific combo. Thx for the lead.
Yes, this is clearly a tradeoff in the approach in that 2-way interviews have the potential to be extended, could even take twice as long. Working on a question/problem together is a great way to establish a working relationship from the start. If it goes really well, we'll always remember it. You gain insights into my thinking process, and I hear yours. This, as opposed to me just asking you questions one-way. For this to work though, maybe a short initial meeting to get to know each other on some basic level and to set the stage for the exercise would be a great way to start.
Ever had the opportunity to try this approach in some way? If so, how did things go?
Great anecdote. I would expect the approach to level-set the relationship from the start. And, in a way, when both sides of the decision to hire take turns asking questions, the participants will find themselves in a better state by cooperating while they may find themselves in a worse state if they do not, which provides prisoner's dilemma outcomes. For example, in a coding interview, I'll go easy on you because I know you'll go easy on me.