Years ago businesses noticed that graduates from Masters of Business Administration (MBA) programs had a toolset that was unusual and useful. These graduate programs were producing people distinct from general undergraduate business programs, in that they had a good understanding of general business principles applicable throughout an organization, combined with analytical techniques that could be applied usefully in a number of situations.
Identifying these graduates as valuable led to an EXPLOSION in MBA training. Huge numbers of ambitious workers went through the programs to increase their earnings. Vast numbers of schools opened MBA programs to sell to these workers and the market got flooded. MBA programs charged enormous fees and could make the case, from existing data, that they would give you a large return on your investment over the remainder of your career. Now that MBAs are plentiful they no longer command the salary premium they once did and have become a more general credential, the lack of which is more likely to hurt you, rather than having it help you.
Anyone who examined the employment market for MBAs should have been able to identify where it was going. As the bubble was expanding, it should have been obvious that flooding the market with a credential would undermine the career premium it was providing. Instead, schools kept increasing enrollments, students kept paying top dollar, and now an MBA is useful, but not the golden ticket it once was.
Machine learning skills are developing along the same path. There’s a massive demand in industry for employees with them, and because of this a huge demand from workers to acquire these skills, and therefore a large number of organizations offering training. Those gaining these skills are seeing improved career prospects and benefits.
A friend of mine was an economics instructor at a small Midwest college. He enrolled in Georgia Tech’s online Master of Science in Analytics and over a number of years took classes and worked his way through. After graduating he had interest from a number of employers and went to work for the Federal Reserve.
The one saving grace that machine learning has over an MBA is that the concepts are quite a bit harder to understand and learn. Getting an MBA is mostly a matter of putting the time in, while machine learning training, in any rigorous program, will be a challenge, and overcoming it says something useful about the person with that credential.
I expect that, despite claims about the depth of demand, the need for machine learning talent is going to peak and decline in coming years. It certainly won’t hurt you to learn about it, if you’re interested, but we’re past the point where it will give you the massive career boost that it did in the past.