How to Build, Coach, and Lead Successful Data Science Teams (Part-I)?


This is a two-part series about my thoughts in 1) building, 2) leading, motivating, coaching, and retaining data science teams. Part-I focuses on building successful data science teams from scratch. As someone who has built teams in academic research settings as well as in the industry, here are a few tips that worked for me and some pitfalls to avoid. Obviously, there is no silver bullet and I am sure readers may have many different opinions / thoughts. Please chime-in in the comments section below.

My premise here is that you already have a leader who has the expertise to interact with the C-suite, senior execs, experts from non-data science domains, marketing and sales leaders, and of course, most importantly-the clients. Granted, this is easier said than done. Nevertheless, assuming you have all that squared away, let’s start! A good data science team should have the following “personas/roles:”

  1. The Modeler and Deep Thinker: This person is your statistics, optimization, and machine-learning Guru who can break down complex business questions and create mathematical models to tackle them. She/he does not need to be an expert coder/programmer. If you do find one who also has these skills, then you better hire them fast.
  2. The Coder: This person is your expert coder in Big Data Analytics algorithms; he/she does not need to understand all the math and stats behind the scenes but must know how to write very, very, efficient code. This person must have handled numerous multi-million / billion record datasets. This person is typically an expert in multiple languages: Python, C++, R, Scala etc.
  3. The Big Data IT Expert: This is your Hadoop / Spark expert with serious data engineering and database skills. He/she should be able to build your Hadoop / Spark cluster if needed, manage it, and ideally have experience in IT architecture and its best practices.
  4. The Visualization Expert: You will need this person to sell your ideas. This person needs to be an expert in Tableau / Watson Analytics / Qlik / Power BI / Kibana etc. He/she needs to be able to translate the results from complex models to visually intuitive charts and graphics.
  5. The Experts-of-all-Trades: This is your Unicorn who does it all!! As the market for data science matures, some Experts-of-all-Trades seem to be sprouting up here and there. Look out for these superstars.
  6. The Enthusiastic Bunch—Potential Future Leaders: These are the indispensable, new college grads from sciences and/or engineering. They are the engines of a good data science team. They can / should be mentored based on their interests to pursue any one of the above roles.

Of course, bear in mind that many times, these personas are like sets in a Venn diagram, but the boundaries are fuzzy. A simple way to check if your team is ‘ready’ is by asking yourself this: “will my team be able to take a data science engagement from data ingestion, data quality assessment, data modeling to successful analytics operationalization?” If the answer is a resounding yes; congratulations, you have your team. Remember, data science is a team-sport. You better have a good team in place before you jump in the deep-end!

Leave a Reply

Please log in using one of these methods to post your comment: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s