In this workshop we will discuss next steps in the development of pathpyG, an Open Source library for GPU-accelerated Graph Learning in Time Series Data on Complex Networks, which is developed at the Chair of Machine Learning for Complex Networks at JMU Würzburg and the Civil and Environmental Engineering Group at Princeton University. Apart from elaborating strategic goals in the development of pathpyG, this meeting will include a hackathon in which we seek to test new solutions for open challenges and implement missing features.