by SHO TSUJI @CogTalesTweet
Most of us that are currently grad students or postdocs have experienced colleagues leaving academia for industry jobs. Even though I am currently a happy scholar, I can very well understand those who venture into industry – be it for making impact on a shorter time-scale and in a more direct manner, for more job security and more regular working hours, or simply for higher pay and the possibility to plan a family and get some savings. And indeed, the fluidity between academia and industry has arguably never been that strong. I find it very important for us young cognitive scientists to know that academia is not a one-way street, and the world outside there is welcoming us warmly, should we choose to enter it.
To begin with, can you tell us a bit about your career during and after academia?
Christine: I finished my Ph.D. in the Psychology Department at Penn in 2016. Towards the end of my Ph.D. program, I did not see myself continuing in academia and participated in Insight, a 7-week data science fellowship for PhDs. And now I work as a data scientist for a private sector company.
Neil: After obtaining my Ph.D. in the Cognitive and Brain Sciences Department at Rochester, I went to the Max-Planck Institute for Psycholinguistics in the Netherlands for a 3-year postdoc. After that, I came back to the US without any concrete plans for life after academia – all I knew was that academia was not what I wanted to continue doing. Now, I am a Science Consulting Coach, meaning that I provide guidance for students and researchers active in academia — and other technical fields – on improving their presentation skills based on findings from cognitive science. I also fold in experience I have from the performing arts, including storytelling and comedy.
You both seemed to be quite sure that academia wasn’t right for you towards the end of your programs – how did your decision process look like?
Christine: It was definitely a rather gradual process of realizing that this was not what I wanted to continue doing. I do really love the slow and deep thought process that is associated with working in academia, but for me, at one point it became too much – my thesis project was really my baby, and I got more attached to it than was probably good for me. Where I am now, I can delve into the methods but stay more objective about the individual projects themselves, especially as things go much more quickly. I also appreciate that I can have a more direct impact now. I was really into the idea of open source projects and sharing growth during my Ph.D. – and somewhere along the line, I started noticing that all these methods I learned for my Ph.D. were not only freely shared and improved upon in online communities like GitHub and kaggle, but also monetizable.
Neil: For me too, it was a more gradual process – in addition to the factors Christine mentioned, I would say that research life is to a large part solitary, at least in my field – you spend a large proportion of time alone in front of your computer. I enjoy talking to people and I find it very rewarding to have a job that is primarily based on interacting.
Once you had made the decision to quit – how easy was it to set foot into a post-academic world?
Neil: For me, the process of finding out what exactly I could do that would fit my interests took some time. By chance, I consulted with friends and colleagues on how to better present their findings – and I found that I enjoyed this a lot since it meant interacting with people, meeting them where they are, and to teach them something new. Also, I already had the relevant knowledge for such a consulting job: Years of experience in academia paired with knowledge on how our brain perceives and processes information. Once the idea was born, the actual practical side was fairly easy: Founding your own company is actually a matter of a few hours of paperwork. (Growing the company is a good bit more work!)
Christine: Participating in the data science fellowship really helped me set foot into an industry. Insight and other fellowships and boot camps are explicitly meant to help scientists transfer into an industry as data scientists. One thing to keep in mind is that if you are reasonably good at statistics and data analysis in academia, you are probably more qualified than you think for a data analyst or data scientist job in industry. The main things most of us needed to learn in these programs were how to find industry-actionable insights and deliver “minimum viable products,” often in the form of web apps or ad hoc projects for bidder companies, and not just interesting analyses. So for instance, I was fairly comfortable with Python and R before but had no idea how to use Flask or AWS or other packages, platforms, and languages relevant for producing web content. But in several intense weeks, you fill in your gaps quite a bit. After the boot camp, the programs help you find a job by linking you with industry partners.
But then – how did you actually succeed in getting a job, Christine, or finding clients, Neil? After all, you had no prior work experience in your respective industries, and I can imagine that your publication list is not of particular interest out there.
Christine: Of course, it’s not that easy. One thing to keep in mind is that the currency “in the wild” is products, not publications. So for instance, in addition to highlighting your skill set in your CV, it might make sense to have a GitHub page to showcase your skills and products by sharing the code and apps you have developed so far.
Neil: For me, social media plays a big role. Having an up-to-date LinkedIn page is always a good idea. And in my case – Twitter. Maybe you could say that Twitter is the prose version of Github? In any case, I started my professional Twitter account at a conference. I tweeted about things I found interesting and relevant and followed people, and I also got the first base of followers. After that, I would tweet articles and pieces of information that would be interesting to potential clients and partners, and thus my Twitter following grew. Nowadays, people get in touch with me via Twitter for an initial conversation about their needs.
It seems like an online presence is indeed crucial. But what about your actual CV? Any pitfalls or special recommendations?
Neil: I guess I’d recommend having some non-academic friends check your CV – as a general rule of thumb, you will want to use more clear language and describe measurable outcomes than for an academic CV to describe your skill set and achievements. Also, build a more standard resume and send that to friends who have never been academics.
Christine: In addition to that, one thing I had to get used to was to be considered an expert on a topic or methodology that I didn’t consider within the purview of my research, but which others saw as “close enough.” But you just have to go with it and learn how to take credit for it. So even if there’s a particular subfield you may never have published in, for people in the industry you could be the expert on that whole field. So it is not wrong to take praise for that kind of thing in your CV if you think it’s interesting for the company you’re applying to.
And can you tell us something about the job interview itself?
Christine: For data science jobs, you usually start with a “behavioral” interview with standard questions, and then later the interviewers show you some data relevant to their company to see whether you can query it, analyze it, and find insights relevant to their business. Sometimes, it can happen that they want you to code “freestyle” on the whiteboard – that can be a bit tricky, since it requires you to have the relevant things in your head. Definitely, something you should practice ahead of a data science interview!
Neil: Groups and individuals I have worked with want to know whom I’ve worked with before and what strategies I’ve used. It helps to have a plan but a flexible one.
Let’s switch to the job itself. I’d be particularly interested in how much you feel like your work life is now determined by others. At least for me, the flexibility I have is really something I value very highly in academia.
Neil: My choice of career was definitely influenced by the same valuation of flexibility. Since I now am essentially my own boss, I am fairly free in when I work – apart from the obvious things like me having to be at a certain place for giving a workshop, for instance. The content of my work is also what I choose and develop myself, so yes, I consider myself still fairly flexible. The only tricky thing is working for multiple employers at the same time, where each of them wants to get out as much as possible simultaneously. Even if I have say an 8h-per-week contract for client A, it is, of course, difficult to work say two hours a day on this client’s work, and then immediately refocus onto client B.
Christine: I would say that there are two types of flexibility in academia: One, you have a flexibility of time – you can often work when and where you want, and you can easily work say 12h for 3 days in a row but then take a day off to make a long weekend. This flexibility is definitely limited in a job where you have to coordinate with people who sometimes get it into their heads to have meetings at 8 am – I definitely suffer at those! But on the other hand, teleworking and flexible hour policies are creeping into even the most old-fashioned of workplaces, and there’s generally more of a premium on getting the work done by a deadline than adhering to a strict 9-5. Work hours vary a lot by a company, and if it’s important to you, you can find data science jobs with very flexible hours. This can be a double-edged sword, however, because I have heard from those who work at companies with no stipulated work hours or vacation limit (work as long and vacation as often as you want!), that you can end up working all the time out of a sense of guilt and/or competition. It’s like academia all over again, if academia for you meant a lopsided work-life balance as it did for me. The second type of flexibility is that of content, or what you do with your time. There’s a really great blog post by a woman who left a neuroscience post-doc for a data science job, and she uses data from a time-tracking app to compare her day-to-day work life before and after the transition. Basically, she spends way less time dicking around with data collection/munging and grants writing, and way more time on data analysis and collaboration (for both self-driven projects and ad hoc requests). This is definitely consistent with my experience so far. Another source of flexibility is that I more often than not receive open questions: “We have this dataset, how can we use it?” A lot of the work is a bit like a consulting job. But this all depends on the type of data scientist you are: there are those that are more on the engineering side of data science, and those, like me, that are more on the analytics side.
What about pressure? Could you say whether it’s more or less than in academia?
Christine: I generally think of there being two ways one can move in business: “managerially” (a function of the people who report to you) and “technically” (a function of your proficiency in some technical content or skill). At some point in academia, they magically decide that because you are technically proficient in a particular field, you must also be able to manage people. Industry keeps these two dimensions uncollapsed. If you want to rise high to become CTO or CEO, or maybe into a higher management position, there can be a lot of pressure. But in contrast to academia, there is more of an option to just stay where you are and get really, really good at one thing, without e.g. being kicked out of a postdoc at some point. Or you can move laterally and collect a whole bunch of skills but never have to manage (many) people. Also note that if you choose to move up managerially in industry, it is likely that you will do less and less direct content work and will have more and more management duties. But if you end up being a professor at a research university in academia, you will end up having to do both – research and management, not to mention teaching.
Neil: Being self-employed is a particular situation. Do I feel pressure? Yes, since I need to generate my own revenue and constantly look out for new clients. But in academia, before you have tenure, I wouldn’t say it’s less pressure. And in the end, I am now convinced of the service I am offering and that I am good at what I am doing, which leads to less pressure than staying in a job I am not convinced I am good at.
That all sounds actually really nice. Is there anything at all that you miss about academia?
Neil: Oh yes, of course. For instance, the resources: Having a travel budget and being able to travel to conferences all over the world; having a good library with all sorts of books and papers available; and last but not least the intellectual resources: You can’t deny that the people around you are on average super smart, and they also know how to do stuff. Anytime you run into a problem you will probably find someone that can help you. Now, I am a one-man show and need to solve my problems on my own. I have various, less formal, networks of resources (including social media communities!) but they’re nothing like being able to walk down the hall to ask a world-renowned expert on a topic.
Christine: I totally agree about the library: sometimes I have to submit requests for papers when I used to be able to read them at the touch of a button. Apart from that, as I said earlier, I did love working on a long-term project that was my own, and I do miss having that time to ruminate. I also kind of miss the writing process, though I do not miss feeling obligated to publish (it interfered with the ruminating). I think for me personally the way it works in the industry makes more sense. You try out an idea this way and that, and if it isn’t working after a while, you move to the next thing. Sure, some good ideas might get lost for lack of that last, desperate experiment or analysis, but this way there is less danger of becoming so obsessed with an idea (to the point that objectivity might suffer) since there is always another project or question to dive into.
Do you have anything else you’d like to share with our readers?
Christine: Although we’ve been talking about us “leaving” academia, in hindsight I can’t quite pin down what it is I left. I’m still using the same methods I did in grad school, albeit perhaps more productively. The fluidity between academia and industry has arguably never been this strong. Think of all the tenured academics surreptitiously borrowed and then outright poached by Uber, Google, etc., and look at industry conferences and see a mix of CTOs, data scientists, and post-docs who present there. I know a good number of academic labs in cognitive and computational neuroscience who are funded by industry as well as scientific grants. You will probably find more commonalities between academia and industry than you’d expect.
Neil: Christine’s points about fluidity are spot on. I’ve worked with corporate and academic researchers and many needs are similar, if not identical. On a different note, I’d like to emphasize that you may feel alone during a transition process, but many support networks and resources exist! Look around online for post-ac and alt-ac groups. I’ve been to in-person meetups for looking over resumes and discussing challenges, I’ve participated in Twitter discussions, and I’ve made deep connections with acquaintances who revealed at some point that they’d also changed paths.
Thank you so much for the interview, Christine, and Neil!
I also want to thank Jurgis Karuza and Ting Qian, who were also part of a panel on exploring job options outside of academia at the Department of Psychology at Penn that this interview was inspired by – some of the statements made in this interview were originally brought up by those two.