4 ms·

A couple thoughts, from an academic who doesn't mind working with industry partners: "In industry, though, no skeptical journal editor will review your work. I

by Fomite 4y ago

A couple thoughts, from an academic who doesn't mind working with industry partners: "In industry, though, no skeptical journal editor will review your work. It’s going to be a bunch of PMs and engineers, who will take you at your word that the analysis is what you say it is." This is at odds with the tremendous amount of consulting I end up doing essentially double-checking industry analysis. I've had at least three projects where I am functionally cast in the role of "skeptical peer reviewer". "They may have critiques, and good ones, but they will tend to stem from domain knowledge rather than a deep familiarity with the statistical properties of an estimator." Similarly, in my work with industry partners, some of the most rigorous methodological discussions I've ever had have been with them. "Writing for an academic audience, rather than plainly and directly for an audience of sharp non-experts" This is definitely true. It's also good advice for folks in academia writing grant proposals to industry. "Academia is characterized by well-trodden problems, hashed over for decades, and negligible novel data for resolving them. Industry is by comparison a mass of green field areas of inquiry with large budgets, minimal bureaucracy, and ample data." None of this has been my experience. Indeed, most of my work with industry has been "How do we get this done on an absolute shoestring budget?" "School tends to train people to inhabit a state where they passively wait for assignments." This is a really weird statement. Any graduate student who has passed their preliminary exams should already be over this. If his friend was a faculty member? It's been years since he had an "assignment". "You don’t need to ask permission to go to the bathroom and you can seize opportunities of your own volition rather than waiting for them to fall into your lap like an essay due a week from Friday." Following on this - this is just a weird view of academia, especially given his example is someone at the faculty level.

majormajor 4y ago
I've seen data scientists who take the "no assignments" idea as an excuse to focus on "interesting" vs "practical" data problems (and IME, like yours, the "no assignments" mindset arises mainly in academia, not industry - I never even finished a MS, but even there saw far fewer academic 'assignments' than industry ones). And if that project is high-budget or has ample data, the stakes get a lot higher. I've seen teams lose their projects and leaders lose their position because the results weren't there. It doesn't matter if they did the math right or wrong, if the results weren't there due to randomness or outside of their control, etc.... lots of people don't want to spend hundreds of thousands to millions a year on a team that hasn't produced working products. The worst were cases where the project had an impact, but a negative one (because of an over-focus on specific areas, say, without getting sufficient full-business context). And often these were cases which some more requirements gathering up front could've avoided, but "you can do all the science and math right and still fail and lose your job" is a particular pitfall that I don't think most academic transfers really appreciate at first.
mattkrause 4y ago
Yup! I always wonder where these caricatures of academia come from. Another thing that stood out to me was "you need to generate good-enough results quickly." I'm sure there are a few lavishly-funded groups where you can spend eons designing the perfect experiment, collecting noiseless data, and writing it up like Shakespeare (Or GRR Martin) . For the rest of us, especially in soft-money jobs, the name of the game is picking one thing to hone for each project and satisficing everything else.
- Fomite 4y ago
  Agreed. I was familiar with "Salami Slicing" well before I had ever heard "Minimum Viable Product"
aqsalose 4y ago
>Similarly, in my work with industry partners, some of the most rigorous methodological discussions I've ever had have been with them. I am willing to believe this depends a lot on which industry you are working in, what are you selling, and who are your clients.

by Fomite 4y ago

majormajor 4y ago

I've seen data scientists who take the "no assignments" idea as an excuse to focus on "interesting" vs "practical" data problems (and IME, like yours, the "no assignments" mindset arises mainly in academia, not industry - I never even finished a MS, but even there saw far fewer academic 'assignments' than industry ones). And if that project is high-budget or has ample data, the stakes get a lot higher. I've seen teams lose their projects and leaders lose their position because the results weren't there. It doesn't matter if they did the math right or wrong, if the results weren't there due to randomness or outside of their control, etc.... lots of people don't want to spend hundreds of thousands to millions a year on a team that hasn't produced working products. The worst were cases where the project had an impact, but a negative one (because of an over-focus on specific areas, say, without getting sufficient full-business context). And often these were cases which some more requirements gathering up front could've avoided, but "you can do all the science and math right and still fail and lose your job" is a particular pitfall that I don't think most academic transfers really appreciate at first.

mattkrause 4y ago

Yup! I always wonder where these caricatures of academia come from. Another thing that stood out to me was "you need to generate good-enough results quickly." I'm sure there are a few lavishly-funded groups where you can spend eons designing the perfect experiment, collecting noiseless data, and writing it up like Shakespeare (Or GRR Martin) . For the rest of us, especially in soft-money jobs, the name of the game is picking one thing to hone for each project and satisficing everything else.

Fomite 4y ago
Agreed. I was familiar with "Salami Slicing" well before I had ever heard "Minimum Viable Product"

aqsalose 4y ago

>Similarly, in my work with industry partners, some of the most rigorous methodological discussions I've ever had have been with them. I am willing to believe this depends a lot on which industry you are working in, what are you selling, and who are your clients.