There is no shortage of jobs for data scientists, but many hiring managers don’t understand the field as well as they should
In the 2020’s rapid-onset recession, few companies are looking for new hires — that is, unless the words “data” and “science” happen to be in the job title.
From being deemed the “sexiest job of the 21st Century” by Harvard Business Review to LinkedIn naming it the “Most Promising Job of 2019,” data science is a victim of its own success. As buzz about the field grew, so did the number of certification programs, job seekers, and positions classified as “data science” jobs.
But even as the number of data scientists grows, those hiring them need to sharpen their skills; data scientists say they aren’t asking the right questions and don’t even seem to understand the field. A “data analyst,” for example, is not a “data scientist,” and it is certainly not a database programmer. Chris Albon, director of machine learning at the Wikimedia Foundationtweeted last month: “I’ve hired many data scientists. I’ve also been hired quite a few times. Every time I read an article about ‘interviewing for data science jobs,’ halfway thorough I wonder what job they are talking about.”

Job listings that miss the mark

Indeed, recruiters struggle with the title, attempting to slap it on every position from systems analyst to visualization specialist.
“Companies are increasingly using the data scientist title for other similar roles such as data analyst or statistician,” Daniel Zhao, Glassdoor senior economist told TechRepublic. “This muddling of job titles is changing the composition of the data scientist workforce and holding down wages as a result.”
“Every time I read an article about ‘interviewing for data science jobs,’ halfway through I wonder what job they are talking about.” — Chris Albon, Wikimedia
Some job listings are simply looking to make a database or programming position sound more exciting. For example, companies often mislabel a “data analyst” position as one for a “data scientist” — a classification error accompanied by a lower salary and a multi-fold increase in applicants according to 2019 research from Glassdoor.
“It’s quite obvious that many data science professionals would be frustrated by companies posting job descriptions that later turn out to be almost a hoax,” explains Harshajit Sarmah in an Analytics India byline warning employers against jargon-filled data science postings. This problem is especially pervasive in India where many data analysts are needed but often marketed as “data scientists.”

Employers don’t understand their own openings

Not only does taking the wrong job lead to poor morale and low job satisfaction, according to Sarmah, but it can degrade a data scientist’s resume. He urges applicants to do as much research on a position as possible to find out specifics about the position before accepting it. “Every company has the right to make the best out of the data science domain; however, that doesn’t mean they would do it at the cost of someone’s career.”
Savvy applicants can see through the jargon and generalizations that plague so many postings; These veterans can quickly tell whether a potential employer knows what they’re talking about — and whether they have any idea of what the “data science” hire will do when they walk through the door — simply by reading a job description.
“[This is the] big problem in the data science industry that I really don’t think we’re taking seriously enough: the vast majority of data science job descriptions do not convey the actual requirements of the position they’re advertising,” said Jeremie Harris, detailing the problem last year in Towards Data Science.
Harris’ post translates “data science job description” into reality, with advice like, “If you meet 50 percent of the [job description’s] requirements, that might be enough. If you meet 70 percent, you’re good to go. If you meet 100 percent, there’s a good chance you’re overqualified.”

They ask for more qualifications than are needed

The co-founder of Sharpest Minds blames recruiters who don’t understand the space and companies who don’t know what they need. He points out that an organization is looking for a data scientist to solve a particular problem, regardless of the technology used to do it.
Companies further decrease their applicant pool when they describe a particular job as requiring too many skills or years of experience according to Mayam Jahanshahi, of TapRecruit, a specialist that uses machine learning to help job posters to accurately describe positions in order to attract the best applicants.
“The thinking is that one of the ways in which you get a good signal-to-noise [(qualified-to-unqualified candidate)] ratio is if you advertise for a more senior role… In fact, we found the number of qualified applicants was lower for the senior data scientist role.”
Jahanshahi told The O’Reilly Data Show podcast that many candidates, especially females, are reluctant to apply for a role if they don’t meet all the qualifications. By overstating the job level, the response is understated.
Truly experienced data scientists and positions worthy of their attention remain an elite group. The practices of the young profession mature exponentially every year, and its management and hiring tactics are simply racing to catch up.

It’s not entirely their fault — the profession is still evolving

The key to attracting the right data science applicants is being as specific as possible about the position’s duties and skill requirements.
As head of data science at Airbnb, Elena Tej Grewal divvied up her 80-person squad into three tracks based on expertise in analytics, algorithms, and inference. She believes specificity about job duties and candidate experience will push data science hiring to mature.
“If we look at another discipline like engineering, there is the helpful shorthand of ‘front end’ and ‘back end’ engineering, which helps you get a sense of someone’s skills or the area of focus. I realize it’s an imperfect distinction, but it gives a better sense of someone’s expertise than simply ‘engineering.’ Data science is still far from this point; this is what we are moving towards.”