For someone thinking about grad school who does not want to go into academia as a career, a PhD advisor with industry experience is not a bad idea. They will understand what you need to succeed in industry and have connections. Might be easier to find in engineering departments.
> PhD advisor with industry experience is not a bad idea. They will understand what you need to succeed in industry and have connections.
My industry experience taught me the following things:
- In industry, there exist quite some deep, interesting (e.g. math, programming) problems that (unluckily) many people in academia don't have on their radar. These kinds of problems often don't fit into the "boxes" of academic disciplines.
- People in industry are not interested that you attempt to work on a breakthrough on some of these problems (even in your free time) - even if this would give the company millions or even billions of money. They will instead actively be fighting you if you question anything non-shallow.
So to answer your implicitly stated question what you need to succeed in industry: keep your mouth shut, question nothing, and shut off your intelligence. Otherwise you are considered to be a troublemaker.
> In industry, there exist quite some deep, interesting (e.g. math, programming) problems that (unluckily) many people in academia don't have on their radar. These kinds of problems often don't fit into the "boxes" of academic disciplines.
I have reasons for being a bit cautious on giving details, but some hints on example areas are:
- Understanding the dynamics (mathematics) of some exotic markets that are currently outside of the focus of investors. Very interesting mathematics is involved, but this is too "mathy" for many economists, and (currently, because the rules of the market dynamic still have to be sufficiently understood) too "vague" for many mathematicians who work in academia.
- If you work on data integration problems and/or LoB business applications of some big, conservative companies, you begin to see that many of these problems are instances of deep abstract mathematical structures that are outside of the focus of the academic mathematicians who work in the respective academic area (think for example into the direction of algebraic geometry or algebraic topology): it is too "applied" for them. On the other hand, people in industry have a hate for people seeing these deep abstract patterns that could simplify the applications.
- If you look deeply into some business calculations, you might think that the mathematics that is used there is "easy". But if thus some "theory" describing the business calculations does not need to describe "complicated" things that (academic) mathematicians love to think about - wouldn't this mean that there could exist a great abstraction that simplifies these business calculations a lot in computer programs?
- If stochastics is used in, say, insurance industry in a much more "simple" way than in probability textbooks: couldn't there exist a much "simpler" (but likely very different) theory of stochastics that is sufficient to describe and solve the kind of questions that people in the respective industry care about (though for sure not the kind of questions that academic mathematicians who work in probability theory care about)? Again: people who work in industry hate employees who think about such questions. On the other hand, people in academia are interested in different questions.
- If you look at the source code of some historically grown (but business-critical) business application, you begin to understand that there is no known data structure documented in academic literature that can describe all these things. So you start deriving it on your own. The problem is: business people don't like such deep thinking about the "correct data structure" for their business problems. On the other hand, academic computer scientists are not interested in this question either, because the things that the data structure describes is deeply intrtwined with how the business processes of the respective company work.
Which industry and in what role and context? From someone I’ve hired to design a CRUD or mobile app, I’m not all that interested in a proof that a particular two-symbol Turing machine happens to do the job. It may be an interesting result from a research perspective, but it’s not what I’m paying for or on the hook to deliver. Instead, it looks like procrastivity that makes my job more difficult.
As a professor, I believe virtually all profs should have industry experience and occasionally go back for a year or two. (I’ve bounced back and forth!)
> As a professor, I believe virtually all profs should have industry experience and occasionally go back for a year or two. (I’ve bounced back and forth!)
For quite some professors I imagine that going back to industry would make them a lot more arrogant. In academia, being surrounded by very smart people dampens the arrogance a lot because you realize that you may be smart, but not that smart. On the other hand, in industry you have much less people around you that can intellectually stand up against you, which easily makes you smug.
How did that affect getting tenure? My experience watching my advisor go through that process is that an industry stint would negatively impact the process.
There are a lot of variables, but from personal experience it also depends on how you talk about that experience in your tenure dossier. I was able to spin a research finding into a commercial product. Due to intracompany politics, that product never shipped. But my tenure committee talked glowingly about my ability to take a research idea and polish it into something that a major software company would pay me to commercialize.
