Back to Careers Blog

Empowering data scientists as a product manager

London-based product manager and Digital Leader 100 finalist, Yetunde shares how launching McKinsey’s first open-source software tool for data scientists and teaming with the best in data science/engineering, made McKinsey an easy choice for her.

I joined McKinsey about two years ago, after completing my MBA at Oxford. Although I was considering several offers, McKinsey had by far the best in learning and development opportunities, so it was an easy choice. The ability to easily tap 4,500 McKinsey technologists across 30+ industry and functional specialties to support my work and growth is just something you can’t get elsewhere.

Empowering data scientists as a product manager
We strive to provide individuals with disabilities equal access to our website. If you would like information about this content we will be happy to work with you. Please email us at: McKinsey_Website_Accessibility@mckinsey.com

I was hired to be a London-based product manager for McKinsey’s QuantumBlack team, which is focused on advanced analytics to help clients make better, more strategic business decisions. Specifically, I was hired to work on Kedro, McKinsey’s first open-source software tool created for data scientists and engineers. It is a Python framework to create data and machine-learning pipelines using software engineering principles.

In my role, I determine how software engineering principles can be implemented into a data science workflow, and how to design and test the user experience. I help my team prioritize deliveries and examine how Kedro fits into the firm’s wider strategy. We also play a role in product marketing, and promote Kedro across the firm, open source community, and globe—such as working on this video:

Video
About Kedro

Building a data science framework

Kedro is the first framework built for data scientists. My work is especially meaningful because I help make it easier for data scientists to share their code and collaborate when they’re building new machine learning products. Nearly 90 percent of machine learning projects don’t make it into production, partly because of the code. Kedro solves that issue, making it easier to prove the value of machine learning. They don’t have to worry about writing production-ready code, and can spend more time building scalable, deployable, and reproducible data pipelines.

For example, when we first launched Kedro, one of our tasks was to improve the Jupyter Notebook by changing the product’s internal architecture. However, the changes made it difficult to extend Kedro, when our goal was to expand the platform’s capabilities to support experimentation. Our team focused on identifying problems with that version by picturing ourselves as the user, mapping out the user journey, and connecting with Kedro users on GitHub and Slack to learn their challenges.

Now, we’re on the third iteration, and we’ve built something users enjoy. It’s incredible to see how we’ve applied design principles to build the best possible product.

Advice for future product managers

Empowering data scientists as a product manager
We strive to provide individuals with disabilities equal access to our website. If you would like information about this content we will be happy to work with you. Please email us at: McKinsey_Website_Accessibility@mckinsey.com

McKinsey offers many product management roles. You could work with internal users, clients, or in a role like mine, which gets to work with both. In a role like mine, it’s important to understand who your users are as well as understand your role.

My role, for example, would be challenging without a technical background. I need to understand prototyping, designing, and developing minimum viable products (MVPs), and Scrum, for example. However, not all roles at McKinsey are as technical as mine. During your interview, ask questions about what you will do, what you need to know and what roles you’ll work with; people here will be happy to answer them.

Breaking down complex, user challenges

Empowering data scientists as a product manager
We strive to provide individuals with disabilities equal access to our website. If you would like information about this content we will be happy to work with you. Please email us at: McKinsey_Website_Accessibility@mckinsey.com

One of the best pieces of advice I’ve received came from a designer on our team: All user problems can be broken down into common, easy-to-understand language. I always keep that in mind. For example, our data engineers kept requesting templating, so we asked them to help us identify the root of the problem.

It turned out that when multiple engineers worked on the same Kedro project, they had trouble running data pipelines on a file, because their usernames were in the file paths of their datasets. Templating was one way to fix the issue, but we needed to understand the core problem to come up with the best possible solution.

I give so much credit to the team who worked on Kedro and am honored to work with brilliant data scientists, software engineers and others who made the project possible.

Find a job like Yetunde's

About Yetunde

Yetunde is a product (solution) manager with McKinsey’s QuantumBlack team in London. Prior to McKinsey, she was based in Johannesburg—working as a data product manager at Absa Group, and a consultant at Engineers Without Borders South Africa and the Awethu Project. Yetunde has a bachelor’s in mechanical engineering and technology management from the University of Pretoria in South Africa, and an MBA from the University of Oxford; and was named a Young Digital Leader of the Year finalist.

For more information on McKinsey's product management career paths, visit mckinsey.com/TechCareers.

Note: The above video was filmed prior to the COVID-19 pandemic.