Continuing on in my series about academia to industry, here’s how I gained the skills I needed to get a job outside of academia.
I mentioned last week that I decided I wanted to leave academia, but then spent a full year preparing and learning what that meant. I started with an understanding that I had lots of skills, but wasn’t sure how those would translate to anything besides academia, and I ended with a job in industry. So I’ll start with the skills you do have that are applicable (yes, even you) and then talk about which ones you should make sure you have before starting to apply:
Skills you have:
Teaching and Presenting: One of the most critical traits desired in people who work with data is that they are able to communicate that data to other people. It is RARE in industry to have someone who can do the analyses AND communicate those to shareholders and relevant decision makers without them scratching their heads in confusion. But the good news: your entire academic career has prepared you for just this event. You know all those times you taught, and tried to get undergraduates to understand what you were talking about? Or those conferences where you build presentations to best present your results? They are training for doing that very thing which is VERY valuable in the world outside of academia. You already have this skill (high fives all around!).
Statistical knowledge: I know I sound like a jerk when I say this, but since moving into industry, I keep having to revise the list of things that I though everyone knew. I was asked “what is a p-value?” my first week, and have been asked more basic questions about statistics than I was ever asked in academia, even teaching undergraduates. You have a career full of experience doing experiments, collecting data, and making sense of that data by running statistical analyses. You are already ahead of the game in this regard.
Working hard: My mother told me: “The beginning of every job is the same, nose to the grindstone and work hard” and she was right. You need to be able to work harder and longer hours than you’re used to. Think back to your PhD – hours like that, only longer and in one place. The good news is that most of the multi-tasking you used to need to do is off your plate for now. The bad news is that means you have 9 + hours a day to focus on one task. Luckily, you’ve been doing this for awhile, and being a workaholic comes naturally to you. And even more luckily, this is not a long term commitment. You need to work hard at the beginning, but I’ve found it rare for people in my company (and other companies my friends have moved into) to work outside of regular business hours. This is a sprint, not a marathon.
Teaching yourself: In industry, they don’t expect you to be ready to go out of the box, like they often do in academia. In most industry environments, there is this whole process called “onboarding,” which I had never heard of before I started my current job. They know it’ll take you a little while to get up to speed. But, as an academic you’ve got a lifetime of experience teaching yourself. Great, it’ll serve you well. You can spend those first few months learning the things you’re going to need to know on the job, and that ability puts you ahead of your peers.
As you can see above, we’re already qualified, and have the potential to be successful. But you do need a few skills, at least for data science, that it is unlikely you have already attained. However, you can pick them up. Here are a few:
Python: Most academics I know code in R. The fancy ones also code in another language (I wrote simulations in C++)(I am not fancy though). While R is an excellent language for data analytics, and arguably the best in the world for statistics (come at me), for getting code into production, you need to know python. It’s similar to R, so shouldn’t be too hard, and if you have experience using R as a object oriented programming language, then it’ll be easier still. I learned python using DataCamp and strongly recommend it as a great resource. It has a jupyter notebook embedded in the site so you’re able to run code as you learn (learning by doing is important for me). But there are also books, youtube videos, coursera and udacity, and probably flash cards that will also help you learn this skill.
Querying database: In academia, we work on what are called “flat” data files, like CSV or excel files. Once you are looking at customer data, a flat file is simply too small. It’s like trying to open a NGS sequencing file – it’s just laughably impossible to do with your computer. So you need to learn how to query a database with schemas, like Postgres or MySQL. It’s a pretty simple programming language, and you’re going to learn to love to join tables. But this is a skill of taking a mountain of data, and finding the flower you want to study amidst the rocks. It takes time and practice, but is learnable.
Docker: I know I’m focusing a lot on the technology, but it has stood out to me as the one thing that’s very different between industry and academia. Docker is a way to run code locally that is entirely repeatable elsewhere. Your code is run within a “container” that is created with all the things you need installed in it, in a requirements file and an image. If you’re missing a requirement that exists locally on your computer, but is not in your file, the requirement won’t run. As a result, with the code base alone, you’re able to entirely reproduce your work, regardless of where you’re running your code from. When putting your work into deployment, this is especially awesome because you know that your container is ready to go before you deploy it. And it’s becoming ubiquitous across industry, so jump on that container bandwagon and download docker.
I’m sure I’m missing some things, and there are wealth of cultural differences that I haven’t even approached (feel free to comment below). But the gist of this post is that you ALREADY HAVE MOST OF THE SKILLS YOU NEED TO SUCCEED IN INDUSTRY. And the ones you lack are trival to learn. Now that we know this,next week we’ll talk about one of the things you need to do to actually apply: a linkedIn profile.