Seismic swarms show the structure Faults responsible for earthquakes are idealized into two dimensions, despite fault zones being complicated, three-dimensional structures. Ross et al. used machine learning to find 22,000 seismic events near Cahuilla, California, during a seismic swarm. They used the locations and sizes of these events to show how the complex structure of the fault interacted with natural fluid injections from below. The authors' methods highlight the complexities of one fault and suggest a way to characterize other faults around the world. Science , this issue p. 1357 Locating 22,000 events from a seismic swarm shows the complex interplay between earthquakes, fluids, and fault geometry. The vibrant evolutionary patterns made by earthquake swarms are incompatible with standard, effectively two-dimensional (2D) models for general fault architecture. We leverage advances in earthquake monitoring with a deep-learning algorithm to image a fault zone hosting a 4-year-long swarm in southern California. We infer that fluids are naturally injected into the fault zone from below and diffuse through strike-parallel channels while triggering earthquakes. A permeability barrier initially limits up-dip swarm migration but ultimately is circumvented. This enables fluid migration within a shallower section of the fault with fundamentally different mechanical properties. Our observations provide high-resolution constraints on the processes by which swarms initiate, grow, and arrest. These findings illustrate how swarm evolution is strongly controlled by 3D variations in fault architecture.