Social Science-ish PhD Pivoting to Industry (Data Science?)

February 19, 2019 9:48 AM

I'm a first-year PhD student studying social variation in acoustic and articulatory phonetics. I'm thinking of pivoting from a career plan centered around getting an academic job into positioning myself for working in industry. I'm wondering what my options are, what material I should be focusing on, and what job prospects are like.

I'm currently at a top research university in California's Bay Area, where I study linguistics, but the quantitative social science subsection of linguistics. I have been considering pivoting toward industry since last year, when there were only three job openings in my field. Additionally, a really good friend in the field had been adjuncting, but she just got a job at a top-five tech company in the area, and I started feeling more excited about that possibility, too.

I've checked out previous questions, but a lot of them have to do with the interview process. So:

I have some experience with R and some personal experience with programming, butI'm looking to extend those skills considerably. So, I have a lot of leeway during my program to take classes outside my discipline, as I have already finished all my required coursework, and I'm wondering what I should focus on that would extend my skills so that they'd be suitable for a job in industry.

The courses I'm thinking of taking are:

+Calculus through Linear Algebra
+The complete computer science core at my university / Algorithms
+Stats 101 / Statistical Methods for Behavioral and Social Sciences / Statistical Learning (Statistics with ML)
+Intro to NLP (in Python) / NLP with Deep Learning

My questions are these:

+What should I be focusing on when taking these courses to prepare myself for a standard entry-level job interview? Is there anything I'd be remiss not to take a look at?
+I understand that the market is hyper-saturated, what with boot camps and dojos and whatever they're calling them now. Do I have a reasonable chance of getting a job?
+I have a good sense of what the career trajectory of a professor is, but less of an idea in industry. I understand that "data science", in those terms, is relatively new, but what might my long-term career look like, including compensation, etc.?

I'm hoping my questions aren't really naive or starry-eyed. I feel I have a realistic outlook on my prospects in academia, but less so for industry, and I just wanted to get a better idea; I'm open to having my dreams shattered, so to speak, but I just imagined I might have an easier time getting a position in industry vs. academia.

posted by lilies.lilies to Work & Money (6 answers total)

Does your university have a data science graduate degree or graduate certification? I would start there as the ciriculum tends to package the courses needed for a basic career foundation in data science and, depending on the specialization you are interested in (machine learning, business intelligence, etc) you can take advanced courses. Linear algebra would definitely help though.
posted by Young Kullervo at 11:01 AM on February 19, 2019

Re: hyper-saturated market - I know this is anecdata, but every single undergraduate that I pre-screened for interviews at Grace Hopper conference last year had data science on their resume (hundreds of people). It's obviously a fast growing field, but seems like the market will be keeping up with demand. In my industry (financial services), you need to have a deep and practical/applied knowledge of the subject, not just knowledge of R and Python.
posted by valeries at 11:39 AM on February 19, 2019 [2 favorites]

With all of these bootcamps and such, there is a lot of concern about the market being saturated.
I'd suggest (and I'm a social science faculty member who has a number of students thinking along the lines of what you're thinking of) to think about how you can portray yourself as different from the rest of the pack.

For sure you need to be taking all of the data science courses that you can, but what makes you special?

To be honest, you might want to talk to some PhD-holding data scientists to see if it is worth it for you to go through this 4-6 year PhD process if your goal is industry. While yes, you're getting paid to take these classes right now and that is awesome, will the PhD make you that much more money or give you that many more opportunities versus just taking the classes and entering the market?
posted by k8t at 11:43 AM on February 19, 2019 [2 favorites]

I think you're in an odd position. No one can really agree what "data science" means, but there's perhaps something of a spectrum between whatever the border between business analytics and data science is and whatever the border between software engineering and data science is. The closer you get to the software engineering side, the more bias there is towards people whose educational backgrounds are CS, if not machine learning specifically. However, linguists are something of a wild card--even if your dissertation is phonetics, whatever incidental computational linguistics knowledge you have sets you apart from someone who did, I don't know, physics. On the other hand, if you're more oriented towards the analytics side of things (which, full disclosure, I'm not), that bias isn't there. Of course, it all varies dramatically with the company--I was on a team far to the software engineering side of the spectrum that was an ex-physicist, an ex-mathematician, someone with a data analytics degree and someone with an undergrad CS degree.

Keeping in mind that my title these days includes the words "software engineer": In terms of career progression, most places I've worked have sort of lumped data scientists in with software engineers, but we had a hard time being promoted because no one knew what "senior data scientist" looked like in the same way they had a sense of what "senior software engineer" looked like. Pay is probably a bit higher than software engineers on average, but probably comparable to other specialised software engineers.

More specifically though, do you know anyone who's a data scientist in your area? Anyone on LinkedIn you could chat with? If you find a company you're interested in working for, maybe try contacting someone there & ask about their career progression.

Me-mail me. My colleagues or I cover a bunch of the career paths you'd be looking at, and I'll be in the Bay Area in a few weeks and would be happy to chat.
posted by hoyland at 5:18 PM on February 19, 2019 [1 favorite]

I am a data scientist in the Bay Area. Looking at your question, I’m kind of wondering if staying in the PhD program is an optimal path for you if you’ve already decided to join industry. You’ll be spending a lot of time in school which you could spend building industry experience. IMHO, you can build the skills on the side and jump to a data job without having to finish a PhD.

To answer your specific questions:
- Yes, the classes you listed seem relevant. You should keep an eye toward possible portfolio projects you could do in these classes that use real data. That way when you interview you have something to point to and show your skills.
- The market in the Bay Area is still good, and will continue to be good as long as the tech industry in general is doing okay.
- Unlike a tenure track professorship, your typical data science career doesn’t have clear cut linear stepping stones. Whether you grow technically, or through management, or sideways to other kinds of engineering is really up to your personal decision making. Salaries vary a lot, from $70k for a data scientist whose really a reporting analyst to >$200k+ for an algorithms researcher at a public company which includes bonuses and stock.
posted by tinymegalo at 5:30 PM on February 19, 2019

Companies like Facebook hire PhD students with your background for Quant User Experience Research intern positions. You could then finish your PhD program with such internships in between which can lead to a full time industry job when you graduate.
posted by joan_holloway at 9:32 PM on February 19, 2019