mlrun

Why we included this project

MLRun is a practical choice when you want machine learning to move from a notebook into production without rebuilding the same plumbing for every project. It orchestrates the whole workflow, from data prep through feature engineering, training, validation, and online serving, as one automated pipeline on Kubernetes. Because it works with your existing IDE and CI/CD setup, data scientists and DevOps engineers can collaborate instead of handing work back and forth across tool boundaries. The automation pays off most for pipelines that must run repeatedly and reliably, and the project covers generative AI workloads alongside classic ML. If Kubernetes is already part of your stack and you want a single place to build and ship ML applications, this is worth a close look.