Pyro is a probabilistic programming language (PPL) built on Python, designed for developing advanced probabilistic models in AI research. It leverages stochastic variational inference algorithms and probability distributions from PyTorch to handle large datasets and high-dimensional models. Pyro also incorporates Poutine, a library of composable building blocks for modifying the behavior of probabilistic programs. The language is expressive, scalable, flexible, and minimal, aiming to balance these principles to support complex models and inference algorithms. Pyro's design allows for arbitrary Python code interaction and supports various inference algorithms, including gradient-based stochastic variational inference (SVI). The project is open-source and has been used to implement state-of-the-art models such as the variational autoencoder (VAE) and the Deep Markov Model (DMM), demonstrating its effectiveness and flexibility.Pyro is a probabilistic programming language (PPL) built on Python, designed for developing advanced probabilistic models in AI research. It leverages stochastic variational inference algorithms and probability distributions from PyTorch to handle large datasets and high-dimensional models. Pyro also incorporates Poutine, a library of composable building blocks for modifying the behavior of probabilistic programs. The language is expressive, scalable, flexible, and minimal, aiming to balance these principles to support complex models and inference algorithms. Pyro's design allows for arbitrary Python code interaction and supports various inference algorithms, including gradient-based stochastic variational inference (SVI). The project is open-source and has been used to implement state-of-the-art models such as the variational autoencoder (VAE) and the Deep Markov Model (DMM), demonstrating its effectiveness and flexibility.