Modeling strong gravitational lenses to quantify the distortions of the background sources and reconstruct the mass density in the foreground lens has traditionally been a major computational challenge. As the quality of gravitational lens images increases, the task of fully exploiting the information they contain becomes computationally and algorithmically more difficult. In order to tackle this challenge, we have trained a Recurrent Inference Machine (RIM) on a large dataset of realistic gravitational lenses simulated from galaxy images from the COSMOS survey and convergence maps from the IllustrisTNG simulation. This dataset contains the dataset used to train the RIM and the Variational AutoEncoder (VAE) used in our work, Pixelated Reconstruction of Gravitational Lenses using Recurrent Inference Machine, presented at the Machine Learning for Astrophysics Workshop at the Thirty-ninth International Conference on Machine Learning (ICML 2022), as well as the checkpoint files for the neural networks. The source code for our work is published under the package name Censai.