We constantly hear about the rising demand for data science skills coupled with the shortage of data scientists to fill this growing need. This disparity seems like a true opportunity for those seeking career positions in the field, yet stories abound describing how tough it is to land a data science job. What’s happening here?
In this article, I’ll discuss some of the reasons why you may not be getting hired for a job in data science. I’m assuming that you have a reasonable academic background, i.e. you have a degree in a relevant field of study (sorry if you’re a French History major as you may feel some pushback), or you’ve “skilled up” by attending one of the growing number of university programs or boot camps, or have completed some MOOC specializations.
Demonstrated Experience Required
Hiring companies typically require demonstrated experience for open positions in data science, therefore you need to stand out in specific ways to gain serious consideration. Here are a few key ways you can stick a feather in your hat to help you land a job as a data scientist:
– Write a paper: Prepare a paper that summarizes your work on a successful project. One of the data science skills that many companies are seeking is
the ability to explain projects in the form of an executive summary which discusses how the work can be used, provides details about the
methodology, and results. The goal is to make your work consumable by a broad audience and for it to be self-explanatory so that other data scientists can build upon it.
– Develop an impressive data visualization: It’s often helpful to use a dazzling visualization to get your audience’s attention before explaining why an analysis or model is important. A well-formed demonstration of your mastery of data visualization techniques can go a long way with hiring managers.
– Build a data pipeline: Prove your competence with building data pipelines, a very important part of the data science process. Show your skills with data access, data cleansing and imputation, data transformation, and model training.
– Spruce up your GitHub: I can’t stress this one enough. Yes, employers look at a candidate’s GitHub repositories. The more you can highlight your past work on GitHub the better. Show off your best code solutions, and make GitHub a one-stop shop for potential employers to use for qualifying you for open positions. Take a look around at the repositories of for other data scientists to get ideas for how to make yours shine.
– Polish your CV for data science: Yes, people still need to have a good curriculum vitae (CV) or resume. Make sure you develop a CV specifically for data science positions. I’ve seen resumes for people applying for data science jobs that have outdated technologies listed (nobody wants to see that you’re an expert with MS-DOS 3.1 or the vi editor). Hiring managers will love seeing that you use the phrase “at scale” a lot.
Obtain and Highlight Contemporary Skills
Many times, companies seeking to fill data science positions will look for candidates to demonstrate how they’ve engaged with the latest technologies. Here is a short list designed to impress:
– Cloud computing: Many hiring companies are looking for data scientists with past experience in cloud computing environments because these platforms provide tools that enable data workflows and predictive models to scale effectively. It’s a good idea to demonstrate your knowledge of various cloud platforms, such as Amazon Web Services (AWS), Google Cloud Platform (GCP), or
– Microsoft Azure: The good news is that many of these platforms provide free tiers for becoming familiar with the platform. For example, AWS has free-tier EC2 instances and free usage of services for low volume requests, GCP offers $300 of free credit to try out most of the platform, and Microsoft offers a variety of free and low-cost options for exploring Azure. With these free options, you won’t be able to work with massive data sets, but you can build experience on these
– Jupyter Notebooks: cutting-edge data scientists are using Jupyter notebooks for their work. Be sure you have a number of sample notebooks available to show off your past projects.
– Deep learning: Deep learning is very hot right now. You should consider tooling-up for this popular field. Become familiar with one of the leading frameworks like TensorFlow, Caffe, Microsoft Cognitive Toolkit (CNTK), PyTorch, MXNet, Keras, or Deeplearning4j. Pick one and run with it so you can really say you’re fluent with a well-known framework.
– Kaggle: If you can muster the time, try your hand at a recent Kaggle data science challenge in order to place relatively high on the leader board. A data science hiring manager may see your participation and ranking as something valuable.
Unwritten Reasons for Failure
Even if you’ve done all the right things as suggested above, you may still encounter roadblocks, and some of them are hidden and mysterious. This article, “What no one will tell you about data science job applications,” provides a great summary. First, due to the use of applicant tracking system software, you may be stopped at the door based on the channel you’re coming in from. For example, the HR person may decide to overlook applications coming in from Indeed based solely
on previous poor performance from that channel. The solution is to come in from other, less frequently used channels. Second, most HR teams don’t know much about data science and therefore have a difficult time judging the merits of candidates. So instead of making a mistake and wasting the engineering team’s time, they just pass on everyone. The solution is to use back channels (like LinkedIn groups, or Stack Overflow data science topic areas) that the engineering
team itself may be monitoring.
Aside from all the tips above, make sure you target companies that are a match for your previous domain experience. If you have a strong background in data science for manufacturers, you might be at a disadvantage when applying for a data science job in healthcare. A glance at recent job postings on popular job portals reveals a requirement for at least 2-3 years of experience in a related field.
Stick with what you know, and you’ll fare much better when seeking a data science position.