Random tip: if a PI publishes over 25 papers/year (more than 2 papers/month), in a heavily data- and optimization-driven research e.g. #MachineLearning in medicine, without releasing their codebase or using public libraries, be HIGHLY skeptical about their results. #phdchat
