Polina Mosolova is a data scientist at SAP with a passion for bringing the full potential of current machine learning research to business applications. She is interested in creative combinations of statistical and machine learning methods for use cases that address real-world problems, and in making sure the resulting models are ones business teams can genuinely understand and act on.

Polina completed an industrial PhD in which she created an applied machine learning framework for churn prediction, enhanced by organizational trust theory and explainable machine learning methods. Her research draws on the ABI framework of ability, benevolence, and integrity to turn model explanations into actionable business interventions, and it examines the practical tensions of delivering both academic rigor and production-ready systems.

Her work on explainable AI covers the differences between interpretability, explainability, and actionable ML, glass-box models such as generalized additive models and Neural Additive Models, and tooling like SHAP. Her broader goal is building trustworthy ML systems where explanations drive decisions and create measurable business value.