The project focusses on training a custom large language model for the legal domain, using textual data that the Jackson Lees Group have amassed throughout years of legal practice. The model is unique in learning the specialised language used in the legal domain, along the way incorporating expert legal knowledge.