Leaving academia - a personal account

I presented my thesis in july, 2021. By contract I was supposed to stay in contract a year more, but a set or bureaucratic circumstances (unthinkable in any other professional field but academia in Spain) left me out by September.

I had been an alright R programmer before, and the reliance of bioinformatics on the Linux ecosystem really helped me consolidate my knowledge of shell programs (including some awk on the side!). I did a couple courses and programs to transition out of academia into either the tech sector or some administrative position within a R&D organism. Here’s what I’ve learned in the process, with the caveat that I’ve been in IT for a very short time so far:

• One would think that research makes you slower delivering, but actually no: not having to deal with organizational overhead means that, if you programmed anything during your PhD, most of the work you can do in a big company will be done in a couple hours a day. What does change is the ammount of meetings and informational and organizational bottlenecks that you can experience. These will be the real drawbacks of your productivity, usually.
• Most people in industry don’t have a clue about what a PhD entails - nor do they care, probably. If the project is big enough, it doesn’t really matter what you did before: business or project experience, in companies, counts more than anything. Sadly, this means lots of directly appliable knowledge (eg: on organizational structures or reproducibility) is wasted if you don’t find an amicable ear somewhere up the ladder.
• While problems in industry can be interesting, so far it hasn’t been my experience, to be honest. In technical terms, they are usually trivial in comparison with most problems at a PhD level (in part because the latter depend on cutting-edge knowledge and one person maintaining some remote library for a very specific use). Instead, it seems project-specific implicit knowledge seems to be rewarded (ie. identifying common errors caused by architectural decisions, for example).
• Most of the computation work in academia is either local development (including lots of optimization for this reason) or, if you are lucky and have an ok budget, clusters. Since I’m in Tech any medium to big project seem to rely on AWS. I find AWS painful at several levels, honestly (I’ll complain about that some other day).
• Reproducibility and documentation takes a backseat way more ofthen than in academia, at least in high-pressure projects with lots of technical debt.

I miss…

• Having more control over projects, their direction and organization.
• The inherent interest of scholar topics.
• The freedom to explore options as you (kind of, sometimes) wish.
• Lots of friends, and many people I found interesting and smart and I looked up to. Great conversations.
• Statistics. I liked statistics and I don’t have a use for them lately :(
• Oddly, writing papers. It was a hard time, but also really good food for thought: how do you make an extremely convoluted analysis approachable to peers who might not be familiar with the method or data?

I DON’T miss…

• The inherent job instability.
• Well… conferences. The whole idea of academic posters, to start with, and the pressure of networking being a key to your (precarious, permanently temporal) job-hunting future.
• The expectations that accompany the job; particularly the broken incentives of academic publishing, or the fact your output is measured in publications.
• The inherent job instability (did I mention this before?).
• Underpaid, long working hours. I didn’t suffer much of the second, luckily, but I also know plenty of people who couldn’t afford to.
• How most good research is centralized over certain centers of resources and ideas - belonging to certain places or labs automatically gives you a halo of respectability that has more to do with funding than talent. That is, research output (read: publications) depends on stable incoming money to fund PhD students, postdocs and labs, which depends mostly on country-specific funding and connections. In turn, this funding produces publications in a feedback loop [1].
• Related to the previous topic: the ocassional smugginess and egos. Less pervasive than people think, but still the case from time to time.
• The inherent job instability (yep very much the main problem I’d say).
• The rounds of peer review were always interesting: even at its worst, I was always curious about how people viewed a paper, and learned much of most reviews. But the rest of the publishing process, the endless grind of sending papers to get them desk-rejected without further explanation, was awful, specially knowing that your posibilities in academia are determined by the artificial scarcity business of academic journals.
• I’ve recalibrated how I feel about travels to conferences. I used to like them, but in hindshight they were unnecessary and not as productive to network as one could think. Also, climate change.

[1]: In fact, Spain has a high ratio of publications to funding precisely because I’d bet that most spanish labs recognize this loop and are trying to beat the game from the other side of the equation, ie, getting european funding through hyper-efficiency to compensate from the not-so-great Spanish funding schemes. Call it excellence, if you are cynical enough.

Any recommendations for finding jobs after academia?

I’m sorry to say I don’t have many ideas, other than the common sense. I was lucky as well: the tech sector happens to be a place where finding jobs is (relatively) easy, bypassing the local 14% unemployment rate. I liked programming, and was more or less familiar with the basics of Python before. Probably other jobs would have been harder to find (eg: R&D or innovation related job posts, I guess - those are hard to come by where I live).

In my experience, most people that I’ve known have some kind of data science master or course, but those kind of data analisys jobs seem to be less common than the regular grinding software development kind (as far as I know), so be prepared for that. The best data scientists also know their statistics, in my experience - my impression is that that does make a difference.

Also, get those AWS/Azure certifications if you can afford it. People seem to value them as it means you are willing to use the common business infrastructures. Alternatively, wait until you get an entry job, and within the first year, get it paid for you by your company (apparently, and unlike in academia, most companies don’t mind spending more money on you if that means you get better at your job).

Other than that, you know what to do be proactive, send not only CVs but also mails expressing interest in what the company does, or even small (<1 min) presentation videos if you really like the place and think you can contribute meaningully to it. Plus, proactively search for jobs before you defend your thesis to get a sense of things, prepare for it formation-wise, and get your LinkedIn in shape (yep, I know, I also hate it but ocassionally it’s useful).

Anyway, if you feel like reaching out drop me a mail and I can add here whatever questions you have. You have it at the bottom of my home page.

This doesn’t apply to my experience!

That’s fine. This is just an opinionated version of my own experience in a very particular set of circumstances (namely a humanities major -> Cognitive Science MA -> genomics PhD -> IT sector, all in Spain) that I wanted to get out of my chest. Your mileage can (and will) vary.