Marysia Winkels is a Lead Data Scientist at GoDataDriven, a leading Dutch data and AI consultancy, where she helps organizations get real value from machine learning. Her technical foundation is a Master’s degree in Artificial Intelligence from the University of Amsterdam, with a focus on data-efficient deep learning, research into how to get the most performance from the least data, an interest that still drives her work today.

That focus shows in her advocacy for data-centric AI. Marysia makes the case that the biggest model gains often come not from collecting more data or tweaking architectures, but from improving the data itself: raising label quality, editing datasets, and rethinking workflows, especially for transfer learning and fine-tuning. She has shared this practical, data-centric approach, including lessons from a data-centric competition, with the community through talks and podcast appearances.

Teaching is the thread that runs through everything Marysia does. She educates as part of her role at GoDataDriven and serves as a co-organizer of PyData Amsterdam and PyData Global, helping thousands of practitioners exchange knowledge. Whether consulting, speaking, or organizing, she is working to make data science both more effective and more welcoming.