Often, machine learning systems must adapt to changing data (e.g., as user behavior or the environment changes). To enable the development of such systems, LinkedIn has released Lambda Learner, a library designing to offer the “stability of offline model training with the responsiveness of nearline incremental training on fresh data.” Lambda Learner enables LinkedIn to add new data to its Generalized Additive Mixed Effects (GAME) models (which power services like job recommendations and Sponsored Content) within minutes. Specifically, with Lambda Learner (now OSS) enables incremental updates to the random effects components of GAME models.