2025 workshop program
Hyrax
The work
Abstract
Over the coming years, numerous large surveys will begin generating large volumes of data, introducing a new scale of data that conventional techniques will be unable to process. To facilitate data discovery, novel tools, particularly machine learning, will be indispensable. One of the challenges associated with machine learning is the substantial infrastructure and tooling necessary for its large-scale operation. In the industrial setting, teams of engineers support machine learning scientists by implementing this infrastructure; however, this approach is clearly impractical in the academic environment. Here we introduce Hyrax - a framework that abstracts away a significant portion of the infrastructure of machine learning, enabling scientists to focus on the science rather than repetitive code.