Given the current state of the economy and more specifically the massive waves of layoffs that are hitting the big tech industry right now, I was curious just what the current job market actually looked like since I haven’t been on the market for over 5 years now. So I started peeking at generic “data scientist” jobs on Indeed in NYC to see what things were like out there. I occasionally wind up giving bits of career/resume advice to people who send me random messages, so it’s worth resetting my internal scales every so often.
Things are… different.
Note: Just to be clear, I’m NOT actively looking for a new job, nor do I think I’m at particular risk for some kind of surprise layoff. No one would tell a low ranking person like myself if they made such decisions anyways.
True to the wisdom that you shouldn’t ask an old dog data scientist like myself about how to become a data scientist, yeesh, my calibration for even what the various levels of data science positions is all out of whack, let alone the breadth of expectations was pretty off.
First, some positive things that jumped at me
The most notable thing is that, finally, there’s some semblance of job ladder in existence — to the extent that job title ladders are meaningful in cross-industry contexts. Searching purely for the term “data scientist” doesn’t yield a ridiculous mishmash of wildly inconsistent employer demands any more. There’s a semblance of a “junior/entry level” position, then there’s a notion of a “senior” DS which seems to be in the 3-5ish year experience range. There’s even occasionally Staff/Principal and Director positions now, which I honestly don’t remember seeing last time I did a serious job search.
The relative sanity of everything is refreshing compared to the wild west of a decade ago where it seemed like every company either wanted a full stack data unicorn or just a fancier data analyst, but just called them the same title. It definitely makes browsing job descriptions less random.
The fact that the Machine Learning Engineer branch of DS coming into its own was also another boon because it seems like a lot of the ML-heavy positions have gravitated into that ladder. It means that I again don’t have to wade through a bunch of job descriptions only to see “Must have 3+ years using Big Framework to build fancypants ML model” and nope out.
Same goes for variations of “Data engineer” roles. Now that the vocabulary to describe the job has been invented, people are using it and it’s making everyone’s life easier.
There’s also better clarity in what companies are looking for now in the job description. Many posts will give a rough sense of whether it’s an infrastructure building job, a modeling job, or a product development job. The number of requests for full stack unicorns that can do everything and own everything has calmed down a ton.
Things that confused me
Figuring out where the heck I am most likely to fit within this newfangled ladder was surprisingly tricky. Being insulated in a giant tech megacorp, we have our own internal leveling system that’s vaguely comparable across other giant tech megacorps because they’re constantly headhunting against each other. But such systems are largely disconnected from what goes on in the rest of the real world. In that internal system, it took me a fair amount of work to get a mere “Senior” tacked onto my title despite work 15 plus years in industry. I’m still a significant amount of work before I can attempt to claim the title prefix of “Staff”. The “Principal” prefix, let alone “Director” are massively out of reach for the foreseeable future.
Meanwhile, in the outside “real world” ladder…. just from years of experience and relevant duties alone, I’d likely fall somewhere in staff/principal/director area. The descriptions sorta fit. Figuring out that difference took a lot of adjustment since prior to joining megacorp I simply held plain, un-decorated titles of “data analyst” and “data engineer” — there weren’t any more advanced titles to really promote into within the org structure.
Things I still don’t like
While the specialization of titles has helped de-muddify the main DS job pool quite a bit, not everyone has shifted to using the alternative job titles yet so there’s still some job description weeding to be done. On top of that there are still some very distinct “flavors” of jobs in the pool that seem to align along industry/domain lines now.
Finance companies are still very often looking for data scientists through their quant-y lens. Still lots of emphasis on modeling, machine learning, writing performant code, and I’m sure they love their l33tcoding related stuff.
Similarly, there’s a bunch of healthcare, clinical research positions that seem to view DS as a kind of support role for medical research, like a statistician or a lab tech. There’s a lot of specialized domain knowledge needed to work in healthcare, clinical trials, and genomics research, so that whole blob of jobs represent a unique cluster I can’t even engage with because I don’t have those skills.
Not surprisingly, the higher you go the more there’s a need to be familiar with not just the vague practice of data science itself, but specific industry domain knowledge and management experience becomes increasingly desirable. Being a generalist (like I tend to consider myself to be) requires some creative pitching to fit into the required slots.
Another thing to dislike is that while a de facto ladder has started to appear, it is FAR from universal. “Senior data scientist” does a LOT of heavy lifting in a number of places. Some of those involve leading teams or doing far-reaching “own all the data things” level of work.
So is the general industry market better now than it was?
At least from a job post level things seem to have improved quite a bit in a short amount of time. The real test would be to go and interview at a bunch of places to see what it feels like to apply and go through the interview gauntlet.
We’re still a very long way from getting laypeople to have any understanding of what a given role does in. But we’ll try.