Deepening our understanding of neutrino properties and how they do (or don't) fit into the expectations from the Standard Model could provide answers to open questions in our universe. Many experiments that aim to measure neutrino oscillations are helping to move us into a precision era of neutrino properties using various neutrino sources and detection methods. IceCube detects hundreds of thousands of atmospheric neutrinos using optical sensors buried in the ice. Its DeepCore subarray extends detection energies into the GeV-scales, where evidence of oscillation is visible. In contrast, the Deep Underground Neutrino Experiment (DUNE) plans to use multiple Liquid Argon Time Projection Chambers, along with various other scintillator detectors, to make extremely precise measurements of oscillations from an accelerator neutrino beam. This talk will discuss the convolutional neural network reconstruction that leveraged the IceCube and DeepCore arrays to constrain the muon neutrino disappearance oscillation parameters. It will also look to the future and the ways to apply machine learning to other, more challenging, multi-detector neutrino projects, such as DUNE.