A system and method for malware detection uses static and dynamic analysis to augment a machine learning model. At the training step, static and dynamic features are extracted from training datasets and used to train a malware classification model. The malware classification model is used to classify unknown files based on verdicts from both static and dynamic models.