Machine learning has played a significant role in High Energy Physics since a long time mostly in event classification at the data analysis level. With advances in data science, and in particular in deep learning, very complex tasks in Internet Of Things are getting solved by learning a solution from data and experimentation. Deep learning has recently also entered Science and High Energy Physics, bringing improved performance at several levels. An overview of deep learning applications deployed in HEP is presented, followed by a broader view on prospects of applying deep learning to multiple challenges in particle physics.