In discussion with an experience researcher and developer, it was suggested that:
"If there was a focus on how one can practically design an AI system while not being bogged down by programming questions, it would bring in and help many people in the community."
As one of the persistent critiques of AI is the relative lack of experienced experts, I think it is incumbent on us to evolve the parameters of this Stack to attract such contributors.
Is "architecture" a sufficient descriptor?
This site was created to be a space for the academic and conceptual and generally not-super-mathy questions that didn't previously have a good home. I think the architecting aspect of "how do I actually do a real-life thing with AI?" (as opposed to "how do I tune this number cruncher?") is worthy of a space, and that it fits here. Conveniently, there isn't a lot of overlap with older sites. Stats.SE is about statistics (surprise!), and while it looks like the setup of systems to solve problems might on-topic at Data Science, that site also seems to focus on the lower-level details.
I would agree that "architecture" seems to describe this aspect well. Architects don't do the manual work of constructing the building, and they certainly don't mix the concrete, rather they make the higher-level plans so the client winds up with a nice house, which is the goal of the whole process.
So, should we expand the scope to include these questions? Personally, I would vote yes.
