Fastcore is the underlying library from which many fast.ai projects are built. It extends the Python language to eliminate boilerplate and provide more convenient, succinct APIs for common tasks. It also borrows ideas from other languages (e.g. Python, Julia, Ruby), such as type dispatch or multiple dispatch, that would be useful in a data science context. Finally, there are utilities, multiprocessing, functional programming, and transforms sections that provide features for various stages of data processing and computation.