Industry or Academia? A CounterpointPosted: August 30, 2013 Filed under: advice, careers | Tags: academia, industrial research labs, research, science 1 Comment
The past several years have seen a few blog posts concerning careers in academia and industry; much of the debate has been colored by high-profile departures of university professors for a home in industry. Notably, various former professors have bemoaned the difficulty of obtaining funding, the difficulty of doing large-scale systems research, quixotic panels and program committees, the lower salaries, and so forth.
Although there is always some truth to the complaints you may have read about, I also feel that these previous posts paint a somewhat one-sided view of the nature of academic life. These posts are often enough to send a graduate student running for industry, thinking that the academic life is not for them. I believe that the picture between industry and academia is a bit more nuanced.
I should provide some context for my comments, which will hopefully also lend some credibility. I’ve now spent about seven years as a university faculty member. I have also worked in at a small startup, large industrial research labs, and at several universities (as an undergraduate, grad student, postdoc, and now as a professor). One of my first jobs was as a programmer for LookSmart, a directory-based search company that was eventually bought by AltaVista; I wrote LookSmart’s first web crawler, back when it was a lot easier to crawl the web. I’ve also spent time at both Hewlett-Packard Labs (where I wrote my first research paper), and I spent a summer at Bell Labs, as well.
Ph.D. students and mid-career researchers alike may wonder about the choice between academia or industry. Here are some of my thoughts on the tradeoffs, based on having spent a reasonable amount of time in both places. First, let’s have a look at some of the things that academia has to offer:
- Freedom. Perhaps one of the most appealing aspects of a job in academia is the freedom that the position offers. A professor sets out the day’s agenda and generally has complete control over how to spend the day. For example, I can determine exactly how much time I want to spend meeting students, writing papers, preparing classes, sleeping, and so forth. I don’t typically have to answer to anyone for how I’m spending my time. This is an extraordinary advantage, because it means that I can allocate my time based on the tasks where I believe that I can derive the most value (and happiness). Of course, this freedom also means that a professor needs to become an expert at time management (more on that in a later post). Another advantage of this flexibility that some professors take advantage of, too, is that they can determine when they want to work; it’s not necessary to work 9-to-5, so if it’s necessary to run an errand during the day, that type of thing is generally completely possible.
- Working with students. One of the biggest differences between industry and academia is the opportunity to work with students, of all stripes. I very much enjoy teaching—one of the main reasons I wanted to become a professor was that I deeply admired my own university professors and wanted to emulate them. Working with students is probably one of the most awesome things about being a professor that is hard to replicate in other jobs. As a classroom instructor, you have the opportunity to interact with a large number of incredibly bright people, who are continually asking questions that shed new light on problems you’ve been teaching for years. Recently, with the opportunity to teach online classes, the opportunity to reach a large number of students has presented itself, which is even more amazing—we now have the opportunity to teach (and influence) an entirely new demographic on a massive scale. As an advisor of Ph.D. students, a professor can shape the professional (and research) development of many students over multiple years. Perhaps even more rewarding is that a professor can learn regularly from students; I have learned many things about a variety of different areas (e.g., statistics, machine learning, software engineering, networking, operating systems) directly from my students; learning from one’s students turns out to be a really exciting (and efficient) way to learn. Working with students and helping them develop research taste, presentation skills, and life skills is simply one of the most rewarding aspects of the job. An industrial researcher may offer that they have the opportunity to work with summer interns—sorry, but it’s not the same. Industrial researchers typically work with a summer intern for a short number of months, not a span of many years (and certainly not continuously over many years). The relationship with the student is not as deep, and (as I have observed many times when my own students take internships) an industrial researcher is typically not as invested in a student’s success or overall career path. There are exceptions—my mentor at one summer internship became a second advisor to me and ultimately became my postdoc advisor, but it stands to reason that the relationship that a professor can have with students cannot be replicated in any other profession, and it is one of the most awesome aspects of the job.
- Easy interaction with experts from other fields. The university makes it incredibly easy to approach experts in other areas. Most universities will not hire an army of experts in a single area, which means that the faculty member down the hall from you (or even right next door) might be an expert in a completely different subject matter. This environment creates an extraordinary opportunity for learning, cross-disciplinary thinking and research, and general enrichment. While it’s always fun to talk to people in your same research area, often the best ideas come from applying ideas across disciplines. The university environment makes this type of interaction much more natural than other environments do.
- Opportunity for long-term impact. As a professor, you do not have quarterly deadlines to meet, monthly reports to file, or a boss that you are regularly accountable for. (In truth, a few funding contracts do have monthly reports, but those are few and far between and generally they are pretty lightweight.) Professors generally have to report to their funding agencies about progress on research projects on an annual basis, but even these reports are typically fairly lightweight and involve describing the research that was carried out over the course of the year—there is very little pressure to meet a specific target or make a particular deadline. Faculty drive themselves to excel (the flip side of this, of course, is that to be a good professor, you really have to have initiative and be a self-starter, because nobody is going to crack the whip on you). Freedom from the constraints of short-term deadlines and having to answer to others who are setting the agenda really presents the possibility of thinking about the “right” solution to a problem (or the right problem to work on), rather than simply hacking together something that just works well enough to get the job done. In my opinion, this aspect of being a professor presents the greatest opportunity for adding value: people working in industry often do not have the time (or interest) in stepping back and thinking about completely new ways of solving a problem, and they may also not be able to bring in completely new perspectives to problem solving. They are often way too busy “fighting fires”. I have found many of my research problems by visiting people who work in industry, asking them about “pain points”, and thinking about solutions to the problems that they face that they simply cannot implement because they don’t have the time to step back and think about a completely new approach.
- Relative stability. It almost goes without saying that tenure is unique to academia; that level of stability offers both peace of mind and the freedom to take risks (something I have taken advantage of; for example, post-tenure, I have taken on much bigger systems-building projects that have taken several years to produce papers). Industry labs are typically not around forever. Many of the industrial research lab “giants” of the past are shadows of their former selves. Even if a research lab does not disappear completely, it can be realigned with a business unit to the point where the “research” becomes advanced development and the employees of that unit lose their relative autonomy. Even in my relatively short research career, I have seen several research labs disappear, often with very little warning. Although one of the oft-stated benefits of working in a research lab is that there is no need to raise funding, the flip side of this feature is that when the funding disappears, you lose the perks that you once enjoyed. Regardless of whether you are in industry or academia, you are always at the mercy of your funders. Personally, I would at least rather hold the pursestrings myself, so that I can determine my own research agenda, the scale of the research program that I would like to pursue, and so forth.
- Bridge-building. A professor is not beholden to any particular person, dogma, ideology, company, or other alliance. This stands in contrast to industrial researchers, who have a mission to contribute to the company’s overall value and also have some amount of loyalty (and ties) to that company. As a professor, I have no such ties. I’ve had meetings with Microsoft in the morning, Google in the afternoon, and Amazon in the evening in a single day. In these meetings, a professor can learn amazing things about industry’s view on problems. It may be difficult for an industrial researcher to get the same perspective: it’s unlikely, for example, that an employee from one company will get to hear about its competitor’s approach to a problem. In contrast, if they seize the opportunities, professors can learn an incredible amount from a diverse body of knowledge and have the opportunity to synthesize that knowledge in ways that others can’t.
- Outreach. Activities that transfer research results to the broader society are part of a professor’s mission. A professor has the opportunity to give tutorials, develop online courses, write blog posts, teach at local high schools, organize and run workshops, and so forth. This is not to say that the same opportunities wouldn’t exist in other research positions, but outreach is often not as central to the mission of other jobs as it is to being a professor. For example, every National Science Foundation grant that I write explicitly asks how I will transfer the results of the research to society; this type of outreach is thus very explicitly part of my mission, and I truly enjoy the opportunity for outreach.
- Summers. Someone once told me that the top four reasons to become a professor are May, June, July, and August. Summer offers huge chunks of unstructured time to pursue larger projects and independent interests. For example, over the past summer, I escaped Georgia Tech for several months, ensconced myself in Cape Town, and used the focused (and relaxed) time away to prepare a massive, open online course on Software Defined Networking, a topic that I had long been yearning to dive into the details on.
I cannot speak as much to the benefits of working in an industrial research lab, as I have less experience, but from my limited experience, I have observed that an industrial research lab has the following benefits:
- Structure. In contrast to a professor’s day, which is completely self-determined, time in an industrial research lab is more structured, typically because interaction with peers (as opposed to students) is more common, and as a researcher you are less of an independent actor. You may have specific team meetings that are set up without your control that occur at the same time every day, for example, or regular meetings with your manager. For some people, this type of structure can be extremely helpful for their productivity.
- Money (when it exists). When times are good, there is no place like a research lab—money flows like honey, and equipment, travel, interns and so forth seem to spring from an infinite well. Unlike professors, industrial researchers do not have to interrupt their rhythm to write grants, visit funding agencies, attend “principal investigator (PI)” meetings, and so forth. Although I personally enjoy the process of writing grants, as it gives me the opportunity to think about the big picture of my research agenda, it certainly does interrupt the flow of work, and it can be distracting, since the scheduling of funding deadlines is beyond my control, and they can certainly come at convenient times. Of course, when money dries up, there is probably no more frustrating place to work than a research lab; when funding disappears to the point where your company won’t send you to a conference to present your paper and you have to travel on your own dime, the once-good life can start to look pretty dire.
- Direct and tangible impact. Professors can often struggle to have their ideas adopted in practice; in fact, I find that I spend a lot of time “evangelizing” my ideas to industry to try to transfer them into practice, where they can have a more lasting impact. Industrial researchers do not face this problem: the problems that that industrial researchers work on are often directly motivated by business units. Although this does remove some autonomy, the positive aspect of the lost autonomy is that the research results are often immediately and directly useful. Researchers at industrial labs such as Microsoft and AT&T can regularly cite examples of the algorithms or methods they invented finding their way into products that thousands (or millions) of people use everyday. It is much more difficult to have that kind of impact as a professor; it is not impossible, but it requires a lot more will, and a lot more initiative.
- Problem focus. Professors often work on many different problems at once. I find this breadth refreshing and invigorating, but it certainly comes at the cost of depth. A researcher in an industrial lab typically has much more time to think deeply about a problem, hack on code, and directly contribute to the solutions that find their way into published papers. Typically, as a professor, the great insights, code, and so forth that find their way into papers are the work of students, not the work of the professor. A professor often simply does not have the time to be as deeply involved in the details of solutions. This, of course, can vary depending on the style of the professor: some professors, for example, advise only one or two students at a time and work very deeply on problems with their students. But, typically, this is not the norm, and even in these cases the student still has more time for focused work. In contrast, an industrial research does have the opportunity to spend many focused hours on a single problem, working with much the same level of focus as a graduate student might.
- Time to do technical work. Professors are distracted with many different tasks—teaching, committees, fundraising, advising, and so forth. Although these activities are often extremely rewarding, they are certainly often non-technical. They are a far cry from hacking on code, for example. Some people appreciate the opportunity to spend countless focused hours on technical work. For a professor, that kind of time is an incredible luxury: it presents itself from time-to-time, but it is certainly not the norm.
Ultimately, the choice between academia and an industrial research lab involves many tradeoffs, and the best “fit” often depends on individual preferences and working style. I personally think that being a professor is perhaps the best job one can have. There are certainly drawbacks, and other jobs certainly offer different benefits and perks that jobs in academia do not offer, but the job is far better than one might otherwise be led to believe from recent posts on academic departures. One can observe that, although there have been several loud departures for industry, there still remains a very large cadre of extremely happy university professors.
[…] Over years I found this dilemma as one of the most frequently asked questions my postgraduate students tend to pose. Although I would slightly dissent on a couple of topics (deadlines, and relative stability) I think this article might really help in making the choice [ARTICLE] […]