luas builds Gaussian processes (GPs) primarily for two-dimensional data sets. It uses different optimizations to make the application of GPs to 2D data sets possible within a reasonable timeframe. The code is implemented using Jax (ascl:2111.002), which helps calculate derivatives of the log-likelihood as well as permitting the code to be easily run on either CPU or GPU. luas can be used with popular inference frameworks such as NumPyro (ascl:2505.005) and PyMC. The package makes it easier to account for systematics correlated across two dimensions in data sets, in addition to being helpful for any other applications (e.g., interpolation).