Academic Profiles for Data Science

If you’re interested in Data Science (DS) as a field and have read enough job postings, you start to pick up a theme: everybody is looking for Academics. One of the three archetypes for data scientists, Academics mix statistical knowledge with machine learning know-how. Many descriptions will also be looking for deep technical expertise (the presence of which in combination with the other two is rare enough to justify unicorn status) or experience in business intelligence and business process integration (even rarer!)

Knowing this, it’s worth spending some time to elucidate which kinds of academic profiles are viable and realistic to expect when hiring for DS positions. As with so many things in the DS world, there are a number of different possibilities and “all of them” is not one of them.

Why Academics?

Let’s take a moment to discuss why a history and training in Academia can be an advantageous component to have on your Data Science team. Academics, in a DS context, will specialize in Predictive Analytics and Machine Learning. On the other hand, they will have less experience in business intelligence, including (the dark art of) creating buy-in across the enterprise. They may or may not be particularly skilled in explaining technical concepts to non-technical people. For Academics with a background in direct education, this may actually be an additional strength.

Knowing this, Academics represent a trade-off between domain expertise and knowledge of the mathematics and technology that back up many of the more intricate solutions that a Data Scientist may be called upon to create. In addition, if your enterprise is in a position to contribute to open source software projects or publish in the literature of your field, the Academic is uniquely suited to that endeavor.

Finally, if you have significant resource constraints around hiring for your DS team and need to make the most bang for your buck, hiring an Academic is usually the first path that comes to mind – they’ll just need a little help navigating the field of business and bringing their creations into reality.

calculator-scientific

Understanding The Background of Your Academics

Academics will tend to come from one of several different backgrounds. Each of these will have a different effect on the composition of your team, and creating diverse teams helps in preventing your data science practice from becoming a monoculture.

Computer Science (CS)

The CS path is one of the first ones that comes to mind when thinking about how Academics gain entree into the field of Data Science. Data Scientists who have an academic background in computer science will tend to have a deep understanding of logic and algorithms. They’ll naturally gravitate towards machine learning applications, but often will also have a non-trivial amount of experience with predictive analytics and statistics. This group will have strong programming skills, which makes them a natural collaborator with technology teams when it comes to implementing their creations.

However, even though they play nice with your developers they may not have as much familiarity with the software development life cycle, and the day-to-day of building software in an enterprise.

Statistics

The statistics path is the other one that immediately comes to mind when it comes to Academics in Data Science. Many of the Academics from this path will gravitate towards predictive analytics, with a secondary emphasis on machine learning. Statisticians may also approach your organization from fields that are strictly outside of Academia, such as the actuarial sciences. Depending on the individual, it may yet be appropriate to categorize an actuary looking to leave the business of insurance as an Academic for the purpose of building a diverse team. The other archetype they might fit is the Quant.

Look for an expertise in either Bayesian or Frequentist statistics (or both!), and experience in experience in statistical and scientific programming languages. Typically, these will be such languages as R, SAS, or Stata. However, they may also have experience in general purpose programming languages.

Mathematics Writ Large

There being so many different fields of mathematics, it is a certainty that you will meet Academics with a background in non-statistical mathematics. They may be looking for a change of pace from academia, or just to apply some of their research and study to real-life situations. Be it Manifold Geometry or Functional Analysis, these individuals will tend to push forward the sophistication of DS practice wherever they are located. This makes them an intriguing target for medium-to-late stage hiring, especially for organizations interested in publishing research. However, they are less appropriate in the early days of your Data Science team.

As a word of warning, there is generally less in the way of scientific computing outside of Computer Science and Statistics. Therefore, Academics with this background may lack some degree of technological sophistication. However, given the easy availability of substantial computing power these days, and the general ease with which languages can be picked up, this will become less of a materialized worry as time goes on.

Other Sciences

It is also a possibility that you will gather applicants from other sciences, such as the various variants of Biology, Chemistry, and Physics. Typically, these will be academics looking for a change of pace. Especially in the fields where grant-based funding is a common and recurring feature of research, having the steady dedication of an enterprise may be a compelling change, especially (hopefully) if it comes with a corresponding improvement in work-life balance. These Academics will also seek to apply their knowledge to your business’ domain in a novel manner.

books-magazines-building-school

Does the Level of Education Matter?

One persistent feature of job descriptions targeting Academics is that they will almost always specify a level of education necessary. In so many hard sciences, the Bachelors degree is only the first step in participating in the field, and further development will almost always require a PhD. Thus, in descriptions that are looking for Academics the proper “Academic” qualifier is the PhD.

However, that is not to say that applicants with only a Bachelor’s degree should be disregarded as candidates for an “Academic” position. Having less investment in their chosen specialty, candidates at the Bachelors level will be more flexible to learn more and new skills, whereas PhD level candidates will have a specific and vested interest in applying their knowledge of their field to business problems in novel ways. It may be the case to think of your Bachelors level candidates as a development prospect in some other Data Science archetype.

On Academics

The saying of “Don’t Judge a Book By Its Cover” rings just as true when it comes to hiring Academics for your DS positions. There are many different frameworks within which Academics will fit, and each has its own strengths and weaknesses. When hiring, it is important to keep in mind what your organization needs. If you’ve got the technological sophistication to implement demanding neural networks in a production environment, but don’t have the know-how to kick it off, you might want to hire a CS or Stats academic. On the other hand, if your organization is in need of some original ideas and fresh blood to really kick things off, looking at other sectors of Academia may net what you’re looking for.

If you haven’t already, read some of the other pieces that I have written on the topic of managing data scientists, or get in touch if you’re interested in exchanging some ideas.

Leave a Reply

Your email address will not be published.

This site uses Akismet to reduce spam. Learn how your comment data is processed.