Description
In this talk, we focus on the importance and benefits of effortless reproducibility as a first-class construct in machine learning infrastructure. From practices around capturing inputs and outputs to code versioning and transitive dependency management.