When is "good enough" not good enough?

The conventional wisdom, and rightly so, is that someone should be put out a baseline or MVP quickly, and that complicated neural nets in the hands of the inexperienced can be huge time sinks.

Yet when I'm a hiring manager, I'm much less enthusiastic for candidates who've *only* done simple MVPs, and haven't shown (for instance) they can customize neural architecture when the task or domain demands it. Especially for senior to staff roles, I'd want someone who can help …

datascience good hiring hiring manager manager mvp neural nets simple