Probabilistic programming languages can be used to describe statistical models such as those used in artificial intelligence and offer automatic methods for inferring model parameters from statistical observations. Compared to traditional learning algorithms, probabilistic languages allow uncertainty to be handled explicitly. They are based on the Bayesian method, which allows a priori beliefs about the distribution of a model's parameters to be refined based on concrete observations.The purpose of this introductory talk is threefold: present the mathematical basics in probability and statistics---needed to understand bayesian learning; hands-on modeling---needed to understand probabilistic programming paradigm; basic inference algorithms---needed to gain an intuitive overview on their implementation.