In this talk, I will present our ongoing work aimed at predicting language recovery after acquired brain injury, particularly after a stroke. Our goal is to predict individualized/personalized recovery trajectories based on algorithms from brain markers, big data and computational modeling. In this regard, I will cover three topics (a) Using machine learning to predict recovery after stroke where we examine neuroimaging scans to extract features that best predict stroke recovery, (b) Using AI to simulate individual patient recovery trajectories after treatment where we have developed algorithms to predict recovery as a function of rehabilitation. In this project we are also examining predictors of treatment outcomes in bilingual aphasia, and (c) using big data mining to create a therapy calculator for aphasia recovery. In the last project we are building algorithms of therapy outcomes based on a large amount of rehabilitation data using a digital therapeutic Constant Therapy. This recent work has potential to provide personalized recommendations for rehabilitation trajectories for individual patients who have suffered from acquired brain injury.