The confused data scientist

So far, data science still one of the hottest fields in the world. And part of what is making it so attractive and valuable is that there is no one definition for the job. But despite all the attention, no one seems to agree on what a data scientist actually is. In practice, data scientists perform many different business functions, work in a variety of departments, and use a seemingly limitless number of tools, platforms, and software package.

This lack of agreement often has big consequences for the companies who are aspiring data scientists. If the two don’t see eye to eye, the data scientist will feel trapped and unfulfilled. Meanwhile, the organization will fail to achieve its business goals.

If you are thinking of becoming a data scientist — or if you already are one, but you are thinking of switching companies — you must figure out your definition of the word. Next, how your potential employer is defining the role. Finally, you need to effectively communicate in order to align those definitions.

A healthy organization should clearly figure out which responsibilities to allocate or assign in order to have obvious titles (such as big data architect, big data engineer or big data scientist ...) to help clarify expectations and coordinate work.

What Does the Company Need?

Some employers may have a very clear idea of their data science needs, while others may not. Here are some probing questions you can ask a potential employer to ensure that you fully understand the expectations of your role.

• What is the current structure of your data science team?

Try to understand how data scientists currently work and how responsibilities are distributed. Small teams may not yet be highly specialized, so anticipate wearing many hats.

• What technical stack does the data science team use?

If some form of SQL is mentioned first, the employer may really be looking for strong product analysts. If dashboarding packages like Tableau are mentioned, you can anticipate BI responsibilities. Python and R imply a stronger technical skill set, and if packages like TensorFlow, Scikit-learn, Spark MLlib or Mahout mentioned you should definitely be familiar with machine learning.

• What department in the company houses the data science team? (For example: Strategy, Engineering, Operations, Product, Finance.)

This can give you a clue to which skill sets and responsibilities are held by the data science team. Data science teams under Strategy willing to solve many different business problems (wearing many hats). Engineering or Operations are often very technical, while those under Finance may have more business intelligence responsibilities. Under product, a data science team may focus more on product analytics and partner with product managers.

• What types of business questions are data scientists currently solving?

The answer to this question may help you grasp whether data scientists are expected to help solve and own business questions or whether they are expected to focus on pipelines and modeling.

Other questions you can ask: 

• How does senior leadership currently work with data?
• Are data scientists embedded with product teams?
• How many data engineers work with the data science team?
• Which functions would I be working with most closely day-to-day?

Hope it helps