Hiring a Data Scientist can be a daunting task, but why exactly is it so difficult? When the Harvard Business Review named “data scientist’ the sexiest job of the 21st century they drew attention to a major change that had happened in the way that businesses view their data. Nowadays it is difficult to find a field that does not have the need for Data Scientist. They are highly sought after in Finance, Healthcare, Technology, Marketing, Ecommerce, Transportation and many more industries, and while there seems to be no common agreement on what a data scientist really is or what he/she needs to do everybody seems to agree that they are necessary for survival.
The Problem with Data Science Hiring
This increase in demand has lead to a number of reactions from those looking for jobs and those preparing them. Universities have started creating Data Science departments and specialized Data Science recruiters can now be found next to every Starbucks and McDonald’s. Thousands of job seekers are taking advantage of this trying to break into the industry, but the sheer volume of topics that are available and all seem crucial at the same time makes it impossible to be both well-rounded and skillful at all of it. This makes a lot of candidates ask themselves, what exactly is it that they actually need to know on the job.
The unfortunate truth is that most companies often don’t know either. What Data Scientists actually end up doing varies greatly, and the lines of what a Data Scientist is are becoming increasingly unclear. This has already lead for more specific job descriptions to emerge just like Machine Learning Engineer, Computer Vision Specialist, NLP Scientist, Deep Learning Specialist and many others. This has made a great mess of Data Science job postings. There are openings for Data Scientists which actually are just asking for a sophisticated Data Analyst with SQL knowledge. There are positions calling for Computer Vision Specialists which actually much closer resemble actual Data Science work, and as a candidate it is almost impossible to make a sense of it all.
How to craft a sensible Recruiting Strategy
Now how can you as a company correctly evaluate your own needs and start looking for the right talent just for you? The answer of this is of course extremely circumstantial, but there are some questions you can ask yourself before going out and looking for the perfect candidate for you:
1. What projects do we want the Data Scientist to work on?
This needs to be the starting point of any major hire like this. If you are going to spend the money necessary to get a top candidate, you need to know what you need them for. Make sure you already know specific the projects they will be working on and what results you want to gain from them. And no, “We have a ton of data in our system and just need to make sense of it” does not count. The Data Science process can be extremely messy and time intensive, and without a clear sense of direction even the best talent will be lost.
2. Do we need a Generalist or Specialist?
This of course depends the kinds of projects you are looking for you new hire to do, but the breadth and depth of the work makes it difficult to narrow down exactly what to prioritize. Specialized or general knowledge? The answer to this depends on the stage of your organization is in. The general idea is that the farther along you are as a organization and the more built out your analytics department is the more you should try to prioritize specialized over general knowledge. Look around and check what talent you already have available and how that matches with the projects you are trying to work on.
3. How many Data Scientists should we hire?
Most companies seem to forget that Data Scientists don’t work in a vacuum. This is not only true on a departmental level, but also inside Data Science teams. They need not only constantly communicate with the ones providing their infrastructure and the end users of their products, but also with each other. The most effective solution tends to be to hire teams rather than individuals. When drafting up your recruiting strategy it is important to set up teams, so that they not only match when it comes to responsibility and technical expertise, but also with personality.
If you are able to answer all two out of the three questions confidently you are already doing better than most companies. These questions are also a great measure of how sophisticated your Data Strategy is, since if you have that thought out in detail hiring the right talent is just a task rather than an existential crisis.
If you are interested in learning more about Data Strategy and Data Science Hiring check out our blog.