vertex-ai-mlops

Why we included this project

Building ML systems on Google Cloud is a lot of moving parts, and this repo covers most of them: 470+ interactive notebooks that follow the full lifecycle instead of stopping at isolated demos. If you're already on Vertex AI, the hard parts are covered in practical detail, from serving models on online endpoints and batch inference to SQL-based paths, feature stores, pipelines, model evaluation, monitoring, and experiment tracking. It also reaches into BigQuery, Dataflow, Dataproc, and Cloud Composer, so you can see how inference and orchestration fit into a broader data stack. Each notebook is a self-contained workflow you can adapt, which makes it work as a learning path for engineers new to MLOps on GCP and as a starting template for production work. The applied GenAI and forecasting sections are also useful if you're moving from custom training into retrieval, evaluation, and time series work on the same platform.