This session will introduce uplift modeling and explain its distinct challenges compared to traditional machine learning approaches. We will explore avoiding bias in training data, choosing evaluation metrics, and how uplift models can optimize competing business trade-offs.
Finally, we’ll delve into the components of an end-to-end uplift modeling system deployed in production, including data pipelines, an A/B testing framework, and methods for serving predictions.
Justin Kaiser is a Senior Machine Learning Engineer at HelloFresh, where he develops personalization systems to enhance millions of customers' meal planning experiences.
