Formulating ‘Out of Memory Kill’ Prediction on the Netflix App as a Machine Learning Problem

The article by Netflix Engineering Team discusses the analysis and prediction of "out of memory" (OOM) kills on the Netflix App as a machine learning problem.

The goal is to take pre-emptive actions to prevent crashes on devices with low memory constraints, thereby improving user experience. The article discusses challenges in dataset curation and labeling, and provides guidelines for exploratory analysis and prediction.

The dataset is comprised of device capabilities and runtime memory data, and is labeled using a sliding window approach. The article also explores some graphical analysis on the structured dataset and suggests feature engineering and accuracy measures for future work.