The recent buzz around Data Science and AI has seen lots of people making career changes into this sector.
If you’re trying to do this while holding down another full-time job, however, it’s easy to become burned out. What starts out as something perfectly manageable (an online course during the evenings) can quickly become overwhelming and, before you know it, you’re writing shopping lists in pandas DataFrames and waking up in a cold sweat humming the StatQuest theme tune.
Trust me, I’ve been there.
Over the last 2 years, I’ve made a career change into Data Science, and, while at times this has been undeniably exhilarating, at other times I have felt completely overwhelmed by the size of the task.
If you’re an aspiring Data Scientist, good on you! You’re on a super exciting path, and I genuinely believe that the world of Data Science is one of the most exciting places to be right now. But be warned – navigating this journey can be incredibly difficult and is likely to put a significant strain on your time.
Through this article, I will share some of my top tips for making a successful career transition while avoiding burnout. If you’re bored of platitudes like "knuckle down" or "take a break" and want to hear a perspective from someone who’s actually done it, this is the piece for you.
What exactly is burnout?
The World Health Organisation (WHO) defines burnout as:
a syndrome conceptualized as resulting from chronic workplace stress that has not been successfully managed. It is characterized by three dimensions:
(1) feelings of energy depletion or exhaustion;
(2) increased mental distance from one’s job, or feelings of negativism or cynicism related to one’s job; and
(3) reduced professional efficacy
If you’re anything like me, it may come as a surprise that the WHO even recognises burnout. But, as the definition above attests, when left unchecked it can pose huge medical and societal problems.
Interestingly, burnout seems to be a problem across all sectors at the moment – the American Psychological Association even reckon that the COVID-19 pandemic sent burnout and stress level to an all-time high. While burnout can be found across all professions, however, there are special reasons why it can be particularly high amongst Data Scientists. And that’s because of the unique way in which many people are entering the field.
Is burnout a problem in Data Science?
When you start meeting other Data Scientists, it doesn’t take long before you start to notice a pattern. Many Data Scientists have gotten into the industry through making intentional career changes, rather than simply having "fallen into it" or studied Data Science during our undergraduate degrees. Take Data Scientist Zeya LT, for instance, who at age 32 gave up a career in policing to pursue Data Science:
I had no mathematics, computer science or programming background, so the learning curve was steep […] I had to juggle between assignments and taking care of a toddler. Remote learning as a result of the COVID-19 pandemic also presented its own set of challenges for me and my family.
Zeya’s story is emblematic of many people’s stories, my own included. For many of us, Data Science wasn’t a career option we knew about when picking university/job options in school. We only came across the field at a later age, and so we’re now trying to make career changes while working full-time in another job or juggling family responsibilities. We do our 9–5, and then have to squeeze in some learning and/or portfolio projects alongside that.
This makes for a pretty intense schedule and creates conditions rife for burnout. It’s easy to end up working late into the evenings or cancelling plans on weekends or holidays. We justify these patterns to ourselves and to our loved ones, saying things like "I need to work on my personal development" or "it’s not really work."
The problem, however, is that while coding courses and personal projects may feel fun in the short term (e.g. on one individual evening), if continually repeated they can gradually become draining. And I mean really draining. What’s sustainable in the short term quickly becomes unsustainable in the medium- to long-term, and your "career change" can morph from a fun personal development activity into a chore that takes you away from the important things in life.
What’s sustainable in the short term quickly becomes unsustainable in the medium- to long-term
Spoiler alert: The risk of burnout never goes away
When you’re starting out on a career change journey, it’s easy to motivate yourself by thinking about the "gold" waiting for you at the end of the rainbow: the fun new career, the salary increase, the word "AI" on your CV. Fixing your eyes on these things helps you push through the proverbial pain and justify spending inordinate amounts of time (and money) on sites like DataQuest and CodeAcademy.
If you’re an aspiring Data Scientist, it might surprise you to hear that this risk of burnout never really goes away, even if you attain the "gold" you originally set out for. The world of Data Science evolves at a meteoric pace, and I can tell you first-hand that there will always be something new to learn and a new job waiting just beyond the horizon, if only you would strive for it.
(At least, that’s how it feels).
Recognising this truth about ourselves is an important first step and it illustrates the problem with the "hustle culture" narrative which tells us to knuckle down and dig our heels in. If there’s always going to be more to learn, then we Data Scientists – whether you’ve landed your first job or not— need to figure out how to approach career development in a sustainable way. We need to figure out how to play the infinite game of this career we’re going for.
Tip #1: Recognise that you can’t (and shouldn’t) learn everything, and focus on the key things
This may come as a shock to you, but you don’t actually need to know everything to be a Data Scientist.
I know, right – shocking.
Unless you’re going to be forming a one-man/one-woman Data Science team, your skills are always going to be complemented by those of others in your company’s broader Data team. And in a team setting, it’s OK if you don’t know how to do something, because the chances are that there will be others who are able to help. Data Science hiring managers know this, and it’s why they don’t require people to know everything before they get the job. Everyone understands that you’ll have to do some learning on the job, so don’t worry about needing to learn everything before you apply to jobs.
Of course, that’s easier said than done, and when I was making my switch I found it really hard to know which skills were "core" and which were just a "nice-to-have." If you’re new to Data Science, it’s easy to end up in "analysis paralysis" where you’re not sure exactly what to learn and end up trying a bit of everything without really committing.
If that’s where you’re at, my advice would be the same as that of Renato Boemer, who made a career change into Data Science in his late 30s:
Choose Python and move on.
Yes, languages like R and Spark and Julia and JavaScript might all have a place in some Data Science teams, but Python is by far and away the most dominant language for Data Science. In my opinion, it’s also the best language for people new to coding, because its syntax and logic are relatively straightforward.
The one thing I’d add to Renato’s advice is that you should probably also learn SQL. It’s been around since 1979 and it’s not going anywhere anytime soon – many large companies have invested time in building data infrastructure based on it and it’s one of the most loved languages by developers. Plus, the nice thing about SQL is that it teaches you how to think about data relationally, which is a very hard-to-explain yet super important cognitive skill when working in Data Science.
Once you’ve got the basics of these languages, start doing some portfolio projects, learn how to store your code on GitHub and "learn by doing." It’s easily the best way to make concepts sink in and it will provide great fodder for interviews and portfolios. If you’re stuck for ideas, take a look at this article I wrote about how to come up with some:
How to Find Unique Data Science Project Ideas That Make Your Portfolio Stand Out
But – and here’s the clincher – that’s all you need to do before you’re ready to apply for your first job. Despite what you might read online, you don’t need to master things like Linear Algebra and Discrete Optimisation before you’re eligible to work as a Data Scientist. Sure, a lot of people who come from mathematical backgrounds did learn those before they got their break in Data Science, but I’m not convinced they’re truly needed for most entry-level jobs.
If you’re not convinced, you might find it helpful to hear that being a Data Scientist in the field of AI/Data is very different than being a Research Scientist in this field. Research Scientists are in many ways closer to being mathematicians and/or software engineers. They’re the ones working at start-ups or Big Tech building out new Data Science tools and algorithms, and as a result they need to have a very deep understanding of the underlying mathematical and engineering concepts. Data Scientists, by contrast, tend to be more on the applied end of the spectrum; the role is much more focused on solving business problems than on developing entirely new technologies and techniques. If you want to see this for yourself, try searching for some Research Scientist job roles and see how they differ from Data Scientist ones.
The point I’m making is that, if you’re looking to become a Data Scientist, it’s OK to not be up-to-date with all of the underlying mathematics or the most cutting-edge technologies and methods. Don’t get me wrong – you still need to have an awareness (you’d look silly if you’d never heard of ChatGPT, or if you didn’t know what a matrix/vector is), but unless you’re being recruited for an NLP role specifically you probably won’t need to be able to describe the architecture of ChatGPT and know the ins-and-outs of LSTMs from day 1. So lessen your burden and don’t worry about learning everything.
Tip #2: Take (at least) a FULL day off each week
This is ancient wisdom, and I think there’s a lot in it.
Even if your current "extracurricular" data project feels fun, it’s really important to press pause once a week and take a full day off to invest in rest. Take a walk, hang out with friends, learn Basket Weaving – the sky’s the limit! Just make sure that you make time for something (or someone) else.
Invest in rest. Seriously.
Why is this so important? Firstly, because when you’re making a career change, it’s easy to be consumed by the "I’ll be happy when…" narrative and forget to enjoy yourself in the present. But here’s the thing I’ve learned throughout making my career change:
Sacrificing relational time will never be worth it.
If you don’t force yourself to make time for friends and family, those will often be the first things to disappear off your schedule whenever the pressure’s on. This was definitely true for me during the early days of making my career change, when I was constantly trying to cram in as much as I could.
Taking a full day off per week was probably the most helpful thing for keeping me sane and keeping the workload manageable while I was up-skilling in Data. It forced me to recognise that my end goal behind the career change was really to create a better life for me and my family (and because I felt like it was part of my calling, but that’s a story for another day), and that helped me realise that it was really, really worth prioritising time with my family in the present because that was the ultimate end goal anyway. Plus, it gave me so much more energy in the rest of the week and helped me keep the workload sustainable over the course of the whole year.
So go on, take the risk! Take (at least) a FULL day off each week. I know that it’s easier said that done, but this practice honestly transformed my journey.
Tip #3: Don’t overdo it with your portfolio
If you’ve read any of my previous content, you’ll have seen that I’m a huge fan of making portfolios, as they’ve played a big role in helping me land my own Data Science jobs.
The thing is, however, you don’t need to go overboard with making a portfolio. Regardless of how fantastic your portfolio is, you’re never going to actually land a job based solely on your past personal projects – you’re going to need to do interviews as well! The purpose of the portfolio is simply to get your foot in the door and give recruiters a flavour of what you can do.
Personally, if I were making a portfolio from scratch now, I would aim for 3–5 projects, and leave it at that. Any more is overkill.
3–5 projects is plenty
One more thing…
And there you have it: my top tips for avoiding burnout during a career change! My hope is that this article helps you learn from my mistakes and expertly navigate this journey that you’re on.
If you liked this article and you’d like to get further tips and insights on working in Data Science, consider following me here on Medium or LinkedIn. Less than 1% of my readers follow me, so it means a lot when you do.
If you’d like to get unlimited access to all of my stories (and the rest of Medium.com), you can sign up via my referral link for $5 per month. It adds no extra cost to you vs. signing up via the general signup page, and helps to support my writing as I get a small commission.
Thanks for reading!
