I am building AstroBind, an open-source, end-to-end pipeline for computing binding-energy distributions of astrophysically relevant species on pure and mixed interstellar ices. AstroBind combines first-principles calculations with machine learning to predict binding energies from geometric and electronic descriptors of adsorbate–surface complexes, enabling rapid parameterisation for astrochemical models. I am extending the dataset beyond amorphous solid water to CO₂, CH₃OH, and mixed ices, as well as N- and S-bearing species, targeting the polar ice inventory revealed by JWST. All models and data will be open-sourced.