3 Key Career Decisions for Junior Data Scientists
Because Data Science is still a relatively new field, it can be tough to know what your career might look like in 10 years' time.
A survey by Kaggle found that people working in Data Science typically spend 1–2 hours per week looking for a new job. But what should we be looking for?
Of course, there's value to be had in simply ‘going with the flow' and not making firm plans. But I think the risk with this approach is that you unintentionally end up on what Paul Millerd calls "the default path," and you start thinking that your only option is to move up the ladder (Junior DS → DS → Senior DS → Lead DS → Head of DS → … → Supreme Lieutenant Commander of DS → … etc.).
In this article, I'm going to outline three key questions which should help bring clarity to your thinking about your career. If you're a new or aspiring Data Scientist, this'll help you think a bit more strategically about your future career and what you actually want to do (rather than just doing what everyone else does).
Question 1: Do you want to be an individual contributor, a manager, or a bit of both?
For Matheus Facure, the decision to leave management and return to being an individual contributor (IC) was a very deliberate one.
Matheus began his Data Science career as an IC at a large fintech firm, and within 3 years had been promoted to a Manager:
The company was growing so fast that almost every IC from my cohort was forced to take on management positions.
While he learned a lot whilst being a manager, Matheus ultimately decided that it wasn't for him (at least, not yet):
I couldn't stop pondering on [sic] all the things I didn't learn as an IC. Three years is very little time to become even moderately good at Data Science […] Even if I am to become a manager again in the future, I feel like I would be a much better one if I had the proper time to first mature as an IC.
Lots of companies are keen to stress that both are valid career paths and offer equal opportunities to progress. Meta, for instance, say "At Meta, we've always viewed being an individual contributor (IC) and being a manager as two equally important and parallel career paths."
But for many, there's a lot of pressure associated with this decision. Take Erin, for example, who worked at Meta:
During a career conversation with my manager, I was asked if I was interested in becoming an IC. I initially thought: "No, I'm not. That's a failure." There is a misconception that if you go from managing to simply contributing, you're dropping down or you're giving up.
The perception is a common one, but it doesn't have to be that way. Deciding whether to be an IC or a manager in Data Science is important because it's about deciding where you'll thrive.
Personally, I like the idea of building a solid foundation as an IC and not trying to skip the important first step of becoming an excellent Data Scientist. If (in the future) I'm ever in a boardroom and my tech team says to me "this will take 3 months of work and cost $200k," I would want to have the expertise to know whether that's reasonable, and the technical fluency needed to earn the respect of my team.
Question 2: Do you want to be a Data Scientist, a Machine Learning Scientist, a Decision Scientist, or none of the above?
Data Science is a fragmenting industry.
10 years ago, being a Data Scientist meant being a jack-of-all-trades, involved in all sorts of business problems from the boardroom to the engine room. In recent years, however, this generalist role has been fragmenting and we've seen the emergence of more specialist roles like Machine Learning Engineer, Analytics Translator, and Decision Scientist.
Each of these roles is still captured under the broad umbrella of ‘Data Science,' but as you're thinking about your future career options, it's worth taking note of the differences. For example:
- Machine Learning Engineers focus more on the implementation/productionising of models. They're closer to software engineers than statisticians.
- Decision Scientists foreground the solving of business problems. They're not just quantitative gurus; they're equally comfortable with Statistics and Psychology, knowing how to navigate tricky problems and solve them with data and ML, but they're less focused on building products or business-as-usual activities.
- Analytics Translators focus on bridging the gap between Data teams and that nebulous entity, ‘the business.' They are experts at scoping out problems and explaining models and data-driven solutions, but are unlikely to spend time coding.
As you're thinking about your future career, it's worth thinking about which skill set(s) you want to develop. However, I've got to add that ‘None of the above' is also a valid answer here. It's OK to still be figuring it out, and Data Science can be a stepping stone into lots of other areas as well, e.g. Data Journalism, Marketing, Product Management, entrepreneurship, etc. My philosophy is that, as long as you're always learning, you'll figure it out, and you can't go wrong.
Personally, this is where I'm at: I'm still figuring out which specialism to go down.
(That's a professional way of saying, "My answer changes every week!").
I'm equally interested in the technical and non-technical aspects of Data Science, and I plan to keep investing in both skillsets. It's a nice way to keep growing while sticking with my long-term vision of becoming a "jack of all trades, master of one".
Question 3: What were you made to do?
In my experience, career conversations can be super boring and sterile.
We get bound up in "strategies" and "exit plans," we let our LinkedIn feeds shape our thoughts about what we should be doing, and we leave little room for joy and curiosity. We frame our Careers as optimisation problems and try to solve for (quasi)quantifiable things like prestige or salary. And who can blame us? We're Data Scientists, after all!
But here's the clincher:
You are not an optimisation problem! You're a unique and unpredictably wonderful individual, and you can't one-hot encode your personality. You have no idea what will happen in the next 1 year, let alone the next 10 or 20, and you can't know in advance what you'll want to be working on 10 years in the future.
You can't one-hot encode your personality
To account for this uncertainty, the final question in my list is: What were you made to do?
Unsurprisingly, this isn't a question for which you can reach a final conclusion at age 21. It's a question you need to repeatedly revisit over the years. Ask yourself: what am I finding interesting at the moment? Which issues get me excited, and which make my blood boil? Note the emphasis there – what gets me excited? Who cares what Joe Bloggs was made to do – what were you made to do?
For example, at the moment, I'm really interested in (1) learning general software development, (2) learning Arabic. When I started, I had no idea how those might marry up with my career in Data Science, but I've since been able to use those skills to build theSQLgym, a bank of SQL practice questions for Data Scientists who want some data to practice with.
I never would have been able to build that kind of thing had I not followed my curiosity in software development, and it's only now that I'm beginning to see new doors open as a result of pursuing that interest.
(I'm still clueless on how/if the Arabic thing might link into my career one day).
Thanks for reading! I hope you enjoyed it and, if you've got any thoughts, let me know! I'd love to continue the conversation.
One more thing –
If you're liking this content, you might want to check out my other projects:
- AI in Five – I run a free weekly newsletter where I share 5 bullet points on the latest AI news, coding tips and career stories for Data Scientists/Analysts, without any "data is the new oil" rubbish.
- theSQLgym – A new site I'm working on. Learn SQL in your browser through 100s of practice questions.
