Thursday, April 7, 2011

Outside the Academy

I recently gave a talk for the mathematics graduate seminar at the University of Georgia, since I was invited to by my old UVA roomie @christopherdrup, who is now a postdoc down there. The title of my talk was the title of this post, and was about my experience switching out of academia.

Unfortunately, some poor planning (why would anybody care to listen to what I have to say?) and verbal diarrhea obscured anything like a point that could have been made. I should have written this post before-hand, to get my thoughts together. Hopefully writing it afterwards will still be useful for somebody.

My point, if I'd focused a little, would have been something along the lines of: If you decide to stay in academia, that's great, and I wish you all the best. If you decide to get out, that's great too, and "real jobs" can be pretty awesome (crazy-better than grad school). This talk may not apply to you now (I wouldn't have listened my first year or two of grad school), but maybe store it somewhere in the background, because things change (I'd have listened two years ago).

I threw some slides together. You're not likely to get much out of just looking at them, since mostly it's just some random pictures. In what follows, I try to recount what I said (or should have, without so much babbling (hopefully)), with some indication about slide transitions.

I began (first content slide, Math) (well, after a quick intro about how I originally wanted to be a math prof, and now have a non-academic job I love), talking about how, yes, I do actually like math. I had a great time in grad school reading about continued fractions, and in the little reading group a few of us formed for number theory. Even at a broad level, the topic of my thesis work (the more involved expression on the slide) was interesting. I told them I wouldn't bother talking about what it was, since it didn't matter, which I think caught some of them by surprise.

However, I never really cared a whole lot about serious math research. I knew being a professor might involve having to do some research, but I was optimistic it could be about math education, or something. I know I never particularly cared about cohomology, or spectra, or spectral sequences, things that my classes for the last few years talked about. (transition on the first slide, to the sketch of the frustrated guy). I failed to mention that it wasn't so much an issue about just being wholly abstract and useless. That had never really bothered me. But I think when it was not only abstract and useless, but also frustrating and not fun, that's when it bothered me.

But that's ok (next slide), because mostly I was in grad school because I wanted to teach at the college level. Getting a phd seems to be a requisite step along the way, so I'd struggle through (a friend, recently, told me this was a stupid reason to go to grad school. I'm not sure about his rationale, besides liking to tell people they're stupid). Teaching was always what got me through. I think I mentioned that they'd need a reason to get through, because grad school is gonna suck a lot sometimes.

Unfortunately, a few years in (transition to the rows of desk), I started not appreciating the teaching gig either. I've come to refer to it as "institutionalized education", and I don't care for it at all. I lost the feeling that school was about the joy of learning, and that, instead, it's all about accreditation. Really a shame, but that's how I see it.

So now it all sucks. I came in liking math, and wanting to teach, and I no longer had either of those. Not a fun place to be. However, grad school wasn't all bad, and I did get a few useful things out of it. (next slide, Running) First up, running. Not much to say here. Physical exercise is probably good for you, and I encourage it, and running just happens to not require anybody else (mostly), and very little in the way of equipment. (next slide, News) Next up, a news habit. In undergrad, I know I was reading slashdot a little (I know this because I remember checking it in the computer lab where I studied abroad), but don't think I read much other news. Sometime in grad school I picked up a multi-hundred feed Reader habit, and it's one I have no intention of trying to break any time soon. I gave them a quick heads up about rss, and was encouraged that a few folks appeared to nod when I asked if anybody followed news this way.

In addition to running and reading and generally avoiding what I was supposed to be doing, I picked up a few side jobs (next slide). Not on this slide are things like: I taught the racquetball gym class one semester, and I worked as a tutor for athletes. But my first odd-job was working with Webwork (online homework system) with Jeff Holt, at UVA. He must have sent out an email to math grads, asking for folks that were interested in tagging webwork problems, so they could be searched and organized. At some point, when Jeff was going to be away for a year, they decided I could be the assistant admin the year before he left, and then the actual admin in his stead. This sounds more impressive than it is, and wasn't particularly taxing. But hey, I figured it'd be useful, because webwork's a cool thing, and I'd use it if I was going to continue teaching. Through Jeff, who has some connections at a publishing company, I got to do some "independent contractor" (I think it was called) work associated with the Rogawski calculus textbook. It started as working with their webwork problem library, debugging problems mostly, but eventually also included editing a chapter (for the second edition), and writing wrong answers for questions, for use with clickers.

More recently (last summer), I worked as an intern in software development at Rosetta Stone. I applied because the then-fiancee of one of the math grads worked there, and she suggested I apply (and acted as an awesome proxy for sending my resume in). I had a great time. I don't remember how much thesis work got done last summer (this is going in to my final year (ended up as final semester) of grad school), but my guess is not a whole lot.

At the end of the summer, I applied with Rosetta Stone to stay on (or come back?), half-time throughout the fall semester while I was still in grad school, and then starting full-time in January when I was done (either with a degree or not, I was pretty fed up with research, and had been for a while - even up in to November I almost told my advisor I was quitting). Also, on the suggestion of @drmathochist, I applied at a little place in Charlottesville, CCRi. I almost cancelled on the interview, since I basically knew Rosetta Stone would work out, and I'd enjoy it, but went anyway, and enjoyed it.

After sort of a crazy week(end) when I was debating between offers, I decided to go with CCRi (next slide, Outside (CCRi logo, boring slide)). I mentioned that maybe it's not the most typical company, but I don't really know about many others. It's a great small place, and the owner is a professor at UVA, so I think it's maybe closer to the academic end of the spectrum than other places. There's a laid back atmosphere, but lots of fun things to work on, and it's full of great people. I should mention that this was also my impression at Rosetta Stone. So while my idea of programming jobs might come from, say, Office Space, The Matrix, or Dilbert, there are other great jobs out there where you don't have to wear a tie or sit in a cubicle.

One of the things it was suggested I might talk about was what my typical day looks like (next slide). My schedule is pretty flexible, so I can get in whenever, and leave whenever, as long as I'm around for my 10am daily meeting (stand in a circle and say what you were up to yesterday, what you're planning for the day - easy, and helps you keep up with the work of others), and long enough overall to get work done. Occasionally I've got a telecon I need to be on, and once a month I need to write a quick progress report. I don't know what a TPS report is, and Lumberg never shows up and asks me to work weekends.

In grad school, you may work on the same problem for weeks or months (years) and not make much noticeable progress most of the time. These days, I do something different every few days. Longer tasks might take (me) a week or two, but I can always see my progress. I get to play with things at all levels, from the math/stats in the background, R for some computations, Java as a big framework, a database as necessary, and Javascript toward the front. While I've maybe dabbled in most of these things, I came in as (and remain) no expert (hopefully I'm improving). But the point is I can learn, and I expect anybody that's bothered starting grad school can say the same thing.

While most of my days are spent programming (which is awesome), other things come in too, to mix it up a bit. The telecons I'm still sorta getting used to (they still stress me out). But I also got to work on a white paper (here's an awesome thing we could do if you gave us money), and that was pretty fun. Every few weeks we might (depending on which project you're on at the time) have a "sprint planning" meeting, where you set some goals for 2-3 weeks out.

I talked a little about the project I'm on (upper-right), where we do threat prediction: given incident reports, where does it seem crime is likely? I also mentioned that I've learned that singular value decomposition is actually useful (I asked if any of them knew this, and didn't notice any hands up), and talked a little about how it's used for text analysis. Why didn't I know this when I was taking linear algebra?

Oh, and we've got a ping-pong table, so days typically involve some of that :)

Ok, fine, some recommendations (final slide). Again, I don't know why anybody would take advice from me. If somebody did, though, this is it: (1) Read. Hook up an rss reader, add feeds for things you are interested in (math or otherwise), and enjoy. (2) Share. Start a blog, show off the things you're doing, and you've got yourself a ready-to-go resume. Even if you've got no formal experience in something, you've got a free way to point people to something so they can see what you can do. I'd hazard a guess that what you can actually do matters more than what grades you got in whatever random classes you took in school to fulfill graduation requirements (especially because I doubt anybody looks that closely at your individual class grades). The flip side of "Share" is "Do". You gotta do something to have something to share. I hated grad school because I never felt like I should be doing the side-projects I wanted to be doing, that instead I should be directing my efforts at my thesis.

(3) Keep Your Eyes Open. Unfortunately, this recommendation could also be "Network", which I always hated hearing at job talks. I'm not a networker. At conferences I'd go and sit through talks, eat the free food, and basically not talk to people. However, look at the jobs I've had and how I got them. Jeff Holt sent out an email asking for people to work on a project, and I did that. From him, I got work with the publishing company. Later on, I put my resume in at Rosetta Stone because a friend suggested it. The same holds for my current job. For not being any good at networking, as I view it from job talks, I have to say that my network has gotten me places (that I want to be!). But since I don't want to call it networking, I'll say Keep your eyes open. Watch for opportunities, and try it if one comes along that looks interesting. You can actually get through grad school and have other jobs and run and read a lot. Of course, if you realize early enough in grad school that you want out, you could skip the grad school bit and just do the fun jobs and the running and the reading...

That's about it for my talk. There were a few questions afterwards. I'm almost certainly mis-remembering them exactly, but I think the general ideas are there...
  • How much programming would you say you need to have to get a job at a place like mine? I'm sure it depends on what other skills you bring to the table, but if you can learn stuff, that's probably the best skill. If you've got a class or two under your belt, and are otherwise smart etc., that's probably sufficient to be optimistic about sending in a resume. If you take up an interest in programming, outside of official classes, blog your learning efforts, and point people at that.
  • Why did you go in to math if you clearly like programming? I've been getting this one since before I started grad school. Apparently others know me better than I do. They were all right, too. But, in my defense, I wanted to teach college classes, and I wanted to teach math, not computer science. I'm not sure why, but I never really had an interest in teaching computer science.
  • There was a question about useful classes (or that I took to be about useful classes, anyway). I don't remember the last useful class I took. I know at least one early on had material I revisited several times, so was probably useful in that regard. None of the classes were (I think) honestly useful for my programming work. I didn't mention (because it didn't occur to me) that I did develop a better appreciation for category theory, and love the viewpoint it allows. And I did recently wonder if "persistent homology" might be useful at work, and certainly some early topology classes would be helpful there (maybe some late ones too?).
  • There was another question I think I didn't answer well, about what else I learned in grad school that was useful, not just course material. I guess what "soft skills"? I said that people always say that having a phd shows you can focus on a hard problem for a long time (and that, in my case, this isn't true, because mostly I focused on not thinking about my thesis). I think the best skill is being able to learn, and I don't think that I got any better at that because of grad school, necessarily. Certainly having the opportunity to teach a course the first time or two focuses your ability to learn. But you get that in your first year or two, and could still get out after a masters and not miss much. Also, being in grad school may have allowed a schedule for me that let me explore other things, which I may not have gotten to do if I'd jumped into a 9-5.
Chris told me my talk was sort of depressing (that's what happens when I talk about grad school, welcome to my world), but that I had some good points (he didn't say what they were, and I didn't ask). He also tells me that another postdoc in the audience found my talk "refreshing" (since other talks don't talk about how grad school sucks). And apparently I said some things one of the grad students in attendance has been thinking. Going in, I figured such a connection with one person was a suitable goal. I hope whoever it was got something useful out of it, besides just knowing that others have felt the same way (if that's useful).

If I'd lost my idealism about teaching in my first year or two of grad school, I hope that I would have had the sense to bail with my masters. I don't think I've picked up anything since then that I couldn't have gotten outside of grad school. I mean, my two big take-aways were running and news reading.

If I'd known, a year ago, that I'd be at a job like the one I now have (or that I was considering applying for them, that they were even out there), I would have started learning more about stats and things, and (hopefully) blogging my efforts, or some other "share"-ing.