I agree. In my experience (PhD in geology actually, and decent modeling/statistics/etc.-type programmer but not 'expert' by HN standards), this sort of dual expertise is quite valuable, especially at a higher level: Not only can you be a liaison between the two domains and, say, proficiently implement an idea or interpret scientific findings in terms of case law for your clients, but you can also see opportunities for solving one domain's problems with another domain's tools, when the problem domain may never even realize there's a good solution (or even much of a problem!).

In terms of management and engineering, dual experience or expertise could definitely help with understanding how to implement features in a way that encourages collaboration or efficiency between devs, or leaves a feature open for extensibility or interaction with potential features that the engineers aren't aware of yet.