Being a Great Data Scientist Manager Has Nothing to Do With Data Science. Here’s Why.

Your team is made up of people, not processors.

Written by Brendan Meyer
Published on Apr. 12, 2022
Brand Studio Logo

To be a great data scientist, you need to be a great problem-solver. You need to be an expert in explaining complex topics, to be dedicated to research, to have the ability to distill technical information and to stay passionate and curious.

Excel in all of the above, and it’s only logical to someday transition into a managerial role. 

But those skills don’t necessarily translate.

“Every new manager learns very quickly that the manager’s role is extremely different from the role of an individual contributor,” Yulia Shchadilova, Kensho’s team lead of machine learning, said. “What made you successful as a data scientist won’t make you a successful team manager.”

So what’s the big difference? Unlike an individual contributor, where you’re frequently tackling technical problems and optimizing the tools at your disposal, a manager is tasked with balancing something much more complex.

People.

“A trap some of us fall into at times is treating the prospect of leading and tasking a team like an optimization problem and trying to optimally load balance your ‘resources,’ as in, your people,” Chris Abriola, STR’s lead engineer, said. “This may seem like the most effective way to solve many problems as quickly as possible, but it ignores the human element and the fact that people perform their best when they’re happy.”

Want to learn more secrets and insight into the skills that data scientists need to develop when they move into a management role? Built In Boston sat down with two managers to find out.


 

colleagues in casual wear sitting at wooden table and having a team huddle in office in kitchen
Shutterstock

 

Image of Chris Abriola
Chris Abriola
Lead Engineer • STR

 

What they do: STR makes the world a safer place by developing technology to solve emerging national security challenges.

 

What appealed to you about managing a data science team?

Leading a team, especially one diverse in personality and technical expertise, is a fun experience. I knew that my ability to form a clear and exciting technical vision would make for a great fit. I’ve had great mentors throughout my academic and professional career who have prepared me for this role. I saw serving as a team lead as the perfect opportunity to give back and be that person for others who are looking to carve their own path. Seeing someone’s face light up when they get a result using skills that you helped them cultivate never gets old.

 

What skills do data scientists need to develop when they move into a management role?

Perhaps the most important thing about stepping into a leadership position is that your colleagues look to you as a leader. For most of us, this begins with demonstrating our own technical competency and having a proven track record of success. It’s crucial that you work hard to develop your own skills and abilities and that you keep doing so as a leader so that your team continues to trust and look to you for guidance and direction.

The progression from being an individual contributor to a leader of others can be a tricky one as you’re no longer responsible for your own time and research initiatives. Once you begin to lead others, you’ll need to strike the right balance between your own personal execution and giving your team members the time and tools they need to thrive. You must make sure that the plan forward is well-defined, and tasks are delegated in a clear and timely fashion. Carving out and divvying up well-scoped tasks requires a great deal of understanding the scientists on your team. People have different interests, styles and expertise, and without a proper alignment of objectives, executing the team’s vision will be much more difficult. All of this highlights the need for effective communication.

Perhaps the most important thing about stepping into a leadership position is that your colleagues look to you as a leader.”

 

What other advice would you give to a data scientist who is managing a team for the first time?

Have fun and remember that your team is made up of people, not processors.

Leading others can be an incredibly rewarding experience, especially when you see the people you’re working with grow into better scientists. The reality is that there will be ups and downs and some bumps along the way. At times, that creates more stress. When those moments happen, keep things positive and look to your team for support to push you through those challenges together.  

 

 

coworkers in a modern office brainstorming while working on laptop and drinking coffee
Shutterstock

 

Image of Yulia Shchadilova
Yulia Shchadilova
Team Lead, Machine Learning • Kensho Technologies

 

What they do: Kensho leverages S&P Global’s data to research, develop and implement AI and machine learning capabilities that drive fact-based, objective decision-making for businesses.

 

What appealed to you about managing a data science team?

The team management role was a natural progression in my career. I was always interested in improving team cohesion and organization. Prior to taking on this role, I was helping my former managers improve the team's workflow. I loved initiating team-building activities and designing swag for the whole team. 

I was very excited when an opportunity to manage a team of machine learning engineers at Kensho came up. I saw it as a chance to continue developing my leadership skills and take responsibility for the team’s growth and success. I was also interested in understanding the bigger picture of the business in my organization.

 

What skills do data scientists need to develop when they move into a management role? 

Your leadership skills will be absolutely crucial on the managerial track. Your ability to listen to people, make decisions and delegate effectively will now play a very important role.

As a people manager, you should learn to listen to what your team has to say and speak last. Anything you say impacts people and may bias others. If you want to know what your team is concerned about, let them talk. Once you listen to everyone, you can act on this information and make decisions. Sometimes, any decision is better than no decision. Some people can’t tolerate uncertainty, so it is important to keep that in mind. Lastly, learn to delegate effectively. There might be moments of temptation to write code yourself because you know how to do it. Instead, coach your team so they know how to do it and feel comfortable taking responsibility.

Your ability to listen to people, make decisions and delegate effectively will now play a very important role.”

 

What other advice would you give to a data scientist who is managing a team for the first time?

Build your new support network early. Your peers may not be as well-suited to advise you as a manager. The type of problems you are facing now won’t be familiar to many individual contributors. Start building relationships with other managers who have been through a similar transition, as they may provide you with practical advice and share knowledge on how their teams operate.

I will never forget how supportive and helpful my new manager was in the first months of transitioning to the management role. He was always available for me to ask questions and find solutions. I also took advantage of the manager workshops and training that my company offers. Learning how to run effective one-on-one meetings and how to manage remote workers was useful at the transition time and beyond.

 

 

Responses have been edited for length and clarity. Images via listed companies and Shutterstock.