Empathy for Data Science Hiring in a Brutal Market

Why landing a Data Science role feels like an uphill battle and what we can do to lighten the load

Introduction

I graduated this week (yay!), and as I’ve been contemplating my journey so far, I thought I’d switch things up from my routine R posts and talk about why things are the way they are in the Data Science hiring landscape (and tech hiring in general).

If you’re in tech and haven’t been living under a rock for the past few years, you likely know that the job market is absolutely terrible. After the hiring boom of 2021–22, it’s gotten progressively worse for various reasons: the COVID aftermath, economic uncertainty (wars, higher interest rates), changing laws, and more recently, AI. I’ve been navigating the US job market since last year, first for summer internships and more recently for full-time roles. And being an international graduate student, believe me when I say I’ve gotten the shorter end of the stick.

The thing is though, it’s not much easier for US citizens either. I know this because I see so many folks complaining on Reddit, X, Blind, and what have you. The overall experience of “getting a job”, with LeetCode-style assessments, recorded video interviews, take-homes, etc. leaves a lot to be desired. In these frustrating times, it’s important to have empathy for the folks on the other side of the table. Note that some of the points I’m going to discuss in this post are inspired by this talk. With that out of the way, let’s dive in!

Where SWE vs DS Roles differ

I started my career as a full-stack software engineer—building UIs, writing APIs, all the good stuff. Before transitioning to Data Science, I believed that once you understand how something works, you can apply it in any context. For example, once you learn React (a tool for building web UIs), you can build interfaces of any shape and size, for any use case, no matter the domain. I learned the hard way that this doesn’t hold up for Data Scientists. Learning how linear regression works and how to apply it to a given dataset is easy, knowing when it’s the right tool for the job and why it’s appropriate? Not so simple. That understanding sometimes requires years of domain knowledge, a deep grasp of the problem space, and intimate familiarity with your data.

So, it’s not like you learn a few algorithms, learn how to manipulate or visualize data, and voilà — you’re a Data Scientist. No, it requires far more. There’s a saying on Reddit that “Data Scientist is not an entry-level job.” Traditionally, people grew into the role of a DS after years of solving problems in their domain. This is why any DS job description on LinkedIn will likely list something like “X years of experience in Y, Z (domains), or a related field” as the first bullet points.

Gone are the days when companies were desperate to hire people with software skills alone. What makes DS hiring even more difficult is that you’re asking so much more from a single person. On average, you expect an entry-level software engineer to write code and have decent communication skills. A data scientist, however, is expected to be a software engineer who also understands the business and domain, communicates effectively, and knows enough statistics and math to not get into trouble. Needing to assess so many skills makes the process harder to get right.

Philosophy on bad hires

Let’s talk about hiring philosophies at two tech giants: Netflix and Google. At Netflix, they hire people fast and pay them really well — but if it doesn’t work out, they don’t fool around: they fire people quickly. Google, on the other hand, is like a walled garden. They make it very difficult to get in, but once you’re in, getting fired is kind of a big deal. It’s not that restructuring or layoffs don’t happen at Google, quite the opposite, but terminations for poor performance are rare and often emotionally fraught. Most companies resemble Google in their fear of firing people: their hiring processes aren’t that tough, but they really want to avoid letting people go once they’re on the payroll.

DS Hiring is tough

If you hire a bad waiter for your restaurant, you might suffer a week of complaints, but you can replace them relatively easily. If you hire a bad data scientist - someone who’s advising on where to spend your marketing budget or building the model that approves or denies loans—you can face serious business risks. A bad hire affects both parties, drags down team performance and morale, and is so feared that hiring filters get dialed way up. Hiring managers, as a result, try to minimize the regret they feel over making a poor hire. Here’s what a People Analytics Partner at a FAANG company said about the final-round decision meeting:

“Wish we could show a radar chart for a candidate about what was evaluated, how advanced each skill category was (and level of certainty) then hiring managers can make an informed decision based on the actual job description instead of looking for perfection every time [...] Right now we can't do that.”

This quote underlines the hiring managers’ mentality: they demand nothing less than perfect. And in this market, where supply heavily outpaces demand, they can.

An AI Arms Race

As AI has gone mainstream, applicants have begun using it during every step in the hiring process. People on the bottom of the food chain can now use AI to tailor their applications, resumes etc. to come off as experts in the field. It has become really difficult to filter people out based on just their resume. People have created AI assistants such as this which can help solve you coding assessments or even provide a quick answer during an interview call. In response, hiring managers have started using AI to filter applicants. This AI arms race has muddied the waters even more.

Conclusion

Data Science hiring right now is a beast of its own, demanding deep domain chops, solid coding skills, and top-notch communication, all wrapped up in an AI-fueled frenzy. It’s rough for everyone, candidates scrambling for that perfect résumé, recruiters chasing unicorns, and both sides stuck in a never-ending loop of assessments.

But if you're trying to break into DS (or stay in), here’s what’s actually helped me and others: focus on building domain expertise, like really understand the business side. Showcase your work publicly. Whether it's blogs, GitHub, dashboards, or open-source contributions, create artifacts that scream, “I know my stuff.” Don’t rely solely on cold applications, they’re mostly a black hole. Instead, reach out, build relationships, and signal genuine interest in specific problems a company is solving.

Finding value in this blog? Please consider subscribing if you haven't already and spread the word to colleagues who might benefit.