Tensors in MLIR are immutable multi-dimensional containers without an assigned memory location. Bufferization is the process of assigning memory buffers to tensors. In this tutorial, we will show how to use the One-Shot Bufferize infrastructure; both from a user's perspective (how to use the pass and how to debug spurious copies/inefficiencies) and from a developer's perspective (making your own ops bufferizable). We will also touch upon related transformations such as “empty tensor elimination”, buffer deallocation” and "buffer hoisting".