My central interest is in applying machine learning tools to solve logical problems, making use of principles from geometric deep learning to utilise logical symmetries. My hope is that by making AI more logical we can create more noise-resistant, explainable and trustworthy systems.