Subscribe to DSC Newsletter

5 Warning Signs that Turn Off Data Science Hiring Managers

As a hiring manager for data analytics positions, I often complain that there are not enough qualified resumes. Most of the resumes that do get passed on to me from recruiters quickly get filed away. Those job candidates belong to one of five high-risk types that I have identified over the years.

These high-risk candidates do not make the cut despite having technical degrees or technical work experience. If your resume falls into one of these types, you should make an effort to remove the risk factors.


1. You have no real-world data experience

For hiring managers, exposure to real-world datasets is a strict requirement. By real-world data, I mean data with missing values, data with duplicates, data with weird distributions, data with typos, data with erroneous time-stamps, and data with no dictionaries. Real-world data require days, weeks and even months of wrestling to tame, after which you inform your boss you are ready to start the analysis.

Did you remember spending weeks at a time cleaning datasets at school? I bet not. If the class is about teaching logistic regression, the professor has no interest in torturing you with dirty data. Many of the public datasets have become so popular that the major problems with the data are well-known (and sometimes corrected), so the best way to train yourself is to find new datasets.

Because so much of an analyst’s time is spent wrestling with data, the hiring manager won’t take a risk with someone who do not have scars from tackling real-world data.

2. You have no business experience

Data analytics is a business discipline. The goal of data analytics is to use data to make better business decisions. Analysts must work in sync with corporate goals. Their first job is to figure out which business problems can be solved better using available data.

People with technical training and technical work experience frequently have no exposure to the business issues. You do not need a full MBA education but to convince hiring managers, you have to demonstrate command of basic business topics, such as cost-benefit analysis, customer life cycle, and so on.

3. You have no accomplishments

How is it possible that your resume contains no accomplishments? You might wonder whether I need new reading glasses to locate the dozens of bullet points.

The truth is most resumes I come across are long lists of job responsibilities with no actual accomplishments, by which I mean the creation of business value. “Use Python to run statistical models to predict customer churn” is a job responsibility, a task that your boss assigned to you. And this same task is assigned to many other people by myriad bosses at myriad companies.

If your resume reads like a job posting, you have a problem.

4. You are in love with one programming language or software

Of course, if we are both in love, then you are golden. But most companies have a history, having invested in certain systems, processes, and software. They may not be the one you are in love with. (Plainly speaking, if you only know R or Python, you are limiting your career options.)

I learned this lesson the hard way. I thought that if someone is willing to invest the time needed to master one programming language, then this person would find it easy to pick up a different language in the same discipline. The reality is the more expertise one has in one language, the less incentive one has to switch to something else.

So if a resume is too focused on one language, and that language isn’t the primary one for my team, I take a pass.

5. You have many short-term data jobs

Data analytics is a craft. This is where theory meets practice. A craft isn’t something one can master in a few months. In fact, it takes up to a few months just to learn about the data, and then another few months to really learn about the data by manipulating and cleaning them. Then, for a project to get executed, resources need to be secured from other teams, after which time is required for development and testing.

So I am very suspicious when a job candidate spends less than one year at an analytics position, and then have multiple bullet points to describe that experience.

If you look at your resume, and find one or more of the above five characteristics, you are striking fear in hiring managers. We are afraid that we have to invest in training and development, and not sure about the pay-off.

To overcome this problem, you should broaden your resume, such as learning about business basics and learning different software. You should also consult career counselors who can help you sharpen your pitch to hiring managers.

Views: 4101


You need to be a member of Data Science Central to add comments!

Join Data Science Central

Comment by Sione Palu on October 28, 2015 at 7:57pm

Yes, for myself I would look at the types of papers or topics they've done.  If I see a CV with Signal Processing, Control System Design, Physics (Astro-Physics, Condense-Matter Physics, Quantum-Electro-Dynamics, etc, etc,...), Image Processing, Mathematics (Differential Calculus, Optimization, Linear Algebra, Numerical Methods, etc,...), Statistics & Machine-Learning, Computer Vision and other domains. There's overlap in the math from these separate disciplines & that's why I prefer not to just focus & put more weight on a CV with a tick to the 5 criteria on this article or on Statistics/Machine-Learning although, they're very important for a data-scientists skill. Such skills signals an ability to learn & understand complex topics in a short period of time. Of course CV's is one thing, but then you have to interview them on skills they put on their CV.  If the CVs lists their academic publications, we already read them prior to interview. We either request copies of those papers or we found pre-print versions of them on the internet (from the candidate's web site or from their co-author's site, or from Arxiv repository, etc,.., ). So, one can basically infer from the candidate's background & his/her publication regarding their abilities.

Comment by Patrick FitzGerald on October 28, 2015 at 5:48pm

Sione, there's probably much potential in non-off-the-shelf applicants - many bank 'quants' started as physicists, I understand:) In practice, do you try to measure such abilities? Or is it mainly a CV thing? Good comment anyway.

Comment by Sione Palu on October 28, 2015 at 1:45pm

The 5 types listed on this article are important but they weigh less to me.  The most important factor in looking for data science candidates is to scan their CV looking for any hints of the ability of fast learning & to understand complex stuff. We had a particle physicist who did her intern at CERN last year with non of the 5 types criteria above on her skills. She was doing monte-carlo simulations while on CERN with no industry nor machine learning experience at all.  That was the ability that signals that the candidate can learn anything & everything pretty fast. She was hired based on that perceived ability.  She's now the star performer in the team. Anyone who has done deep level study of Quantum Mechanics like this CERN physicist, can pretty much learn anything on the fly. IMO, I think recruiters are short-sighted in what they're looking for, where potential abilities is overlooked instead of looking to tick if the 5 criteria listed on this blog is met or not.

Comment by Nishant Srivastava on October 27, 2015 at 6:18pm

All, I am pleasantly surprised by so many comments and discussions about this article. Full disclosure. I am no blogger. We are trying to solve the talent problem in the field of data science and analytics. We will be posting more such articles on talent in this field.

Check us out on or Follow us on

Responding to Patrick's question: Being a hiring manager for over a decade at a large organization,in general anything less than 7-8 years of experience should be no more than 2 pages. The point in the article was more about quality than quantity. 

Comment by Michael Tamillow on October 27, 2015 at 6:08pm

So, I'm guessing you are in the no accomplishments category, since, well, you know, you have one blog post and you are obviously an expert.

Comment by Patrick FitzGerald on October 27, 2015 at 6:00pm

Thanks for your article, Nishant, very interesting. One question that springs to mind is: how much detail does an applicant provide without writing a book. For example, one sees the caveat 'in not more than two pages'.

Comment by Babaji Grove on October 27, 2015 at 10:07am
Great article! As a recruiting professional I would add career progression, validated research work/publishings, pedigree of work experience and pedigree of academic experience. These are all things one can review via the resume which seems to be a condition. If to look at interviewing said candidates that opens a new realm of considerations. Thanks for sharing :)
Comment by Joell Williamson on October 26, 2015 at 9:17am

@Russell - I have often found that if you can measure the issues, you can show progress and accomplishments.  @Kening - there is data all over the web for use in learning. Find some in your chosen field and start looking at it.

Comment by Russell Greenberg on October 26, 2015 at 8:05am

Great Posting. There is no getting around the basics you list here. Unfortunately, you get the most "recognition" by executing leading edge algorithms, but that is generally the easiest part of the job.

I do express my gratitude to those who develop and release the algorithms. Hat's off and keep'em coming.

Comment by Kening Ren on October 26, 2015 at 5:04am
Oops. Typo: career. Thanks.

Follow Us


  • Add Videos
  • View All


© 2017   Data Science Central   Powered by

Badges  |  Report an Issue  |  Privacy Policy  |  Terms of Service