Making Sense of Confusing Data Science Job Postings
If you’re searching for your first or next data science job, the job placement ads probably vary from confusing to downright laughable (“Must have 10+ years of deep learning experience…”). It’s not that you’ve got to be some superhuman data scientist proficient in deep learning algorithms before deep learning was possible. You will, however, need to get familiar with the reasons data science job postings can be so incoherent.
Recruiters Aren’t Experts
One of the biggest reasons job postings can be ridiculous is the person placing the job add has no idea what the job actually is. Highly technical positions in new industries don’t have precedent, so a recruiter plugging in fields may be fitting a square peg into a birdcage, i.e., job descriptions that have absolutely nothing to do with what a business actually needs.
Many non-technical jobs do require a heavy load of experience, but data science isn’t like that. Not exactly. While experience does sharpen your skills, how do you measure for innovative thinking within deep learning algorithms? How do you quantify an understanding of privacy best practices when dealing with big data?
The answer is you don’t. If you come across the posting that just doesn’t make sense, cut the recruiter some slack. Here are a few things you can do:
- Pull out keywords. 10 years in deep learning may be ridiculous, but do you know anything about deep learning? Showcase your expertise and the time probably won’t be a huge issue.
- Build a theme. The job posting sounds a lot like architecture rather than data science, so maybe they’re looking for a data engineer. Arrange your resume and portfolio accordingly.
- Ignore job titles. Many of these titles are used interchangeably by those who aren’t as familiar with the industry. Until the industry gains consistency, you may want to consider all job titles including data scientist, data engineer, or data analyst despite their differences.
Businesses Don’t Know What They Need
All businesses know what they want, a data science initiative. However, what an organization actually needs is less defined. As teams feel the pressure to launch tech initiatives to stay competitive, they may overgeneralize the job description to make sure they’re getting every possible scenario covered. Most of them won’t need this kind of generalization.
[Related article: Here’s Why You Can’t Hire a Data Scientist]
When you’re looking at these postings here are a few things you can do to make more sense of what’s going on.
- Consider the field. A data science job posting in healthcare could be very different from one on manufacturing. While healthcare may be looking at predictive trends in health insurance, manufacturing may be monitoring machines. Do a little research on some burning questions within the field, and that could give you a better idea of how to approach the job posting.
- Adapt to the business. An enterprise working with data might have different needs than a startup. Brick and mortar based business is not the same thing as a Silicon Valley SaaS. The type of business may be a clue to what your duties are more likely to be.
- Pay attention to the soft skills mentioned. Most companies aren’t going to care so much about the languages. They want to know you can solve problems. Clues such as “driven” or “thrives under pressure” could tell you a lot about the environment you’re heading into.
Increasing Your Chances of Getting Hired
Daniel Gutierrez of ODSC has some advice for Data Scientists trying to land their next job. Not only are job postings sometimes challenging to follow, but you may still have issues if you’ve neglected the one thing you need to demonstrate as a Data Scientist, expertise.
Even when jobs don’t have a clue how much experience they need, you must have a way to showcase that you do know what you’re talking about. A blank GitHub profile, no published papers, no portfolio (even a student one) and your chances of getting noticed are practically zero.
Businesses want to know that you can deliver value, not just that you know your way around TensorFlow or PyTorch. The language doesn’t matter as much as the answer. The one critical function you serve on someone’s data team no matter the job title is delivering business value.
So spruce up that portfolio and get your GitHub profile humming along. A company will go to great lengths to find someone who can demonstrate not only knowledge expertise but an innate sense of how to translate that expertise in a business or organizational setting.
[Related article: 7 Reasons to Start (or Update) Your Github Profile]
Letting It Go
In some cases, the company posting the job description is looking for a unicorn, and you’ll need to decide if you want to sort that out. If you’re up for putting your data science team, pipeline, and compliance together on your own, go for it. If you’re up for adapting the needs and description of the job over time as the company figures out what it needs, more power.
However, some job postings may be ones you just have to let go. The risks involved with an organization that’s dangerously out of touch with best practices and what needs are may be more than what you bargained for.
Otherwise, a little thought into how you might sort out what your real job title is by pulling keywords or building a thematic understanding of the job description based on what you know about the field and organization could help you wade through some seriously illogical postings.
Let us know in the comments some strange job postings you encountered and how you handled them!
Read more data science articles on OpenDataScience.com, including tutorials and guides from beginner to advanced levels! Subscribe to our weekly newsletter here and receive the latest news every Thursday.
