When I decide I need to hire someone, I seek out applicants on my own, based on reputation in industry, published works, patents, and other factors. When I identify someone I want to hire, I send my talent team to make contact in person (i.e. stalk them haha), often literally with a tap on the shoulder.
Who do you think work for? The CIA?
I would say you can define most of machine learning as statistical function approximation. I would also add that the theoretical justifications are indeed important and most good texts find some way to justify the computational methods described.
That said, there are many methods in the literature that are heuristic in nature and for which there are no obvious theoretical justifications.
Failure is more frequently from want of energy than want of capital.