Serving MLOps
noun
Definition: The process of making a trained model available to generate predictions via online or offline inference [Google Machine Learning Glossary].
Example in context: “The new service was implemented to cover a large variety of ML use cases, therefore, some of the main features were, capability for the users to maintain multiple notebook servers, provide a common workflow to perform data retrieval followed by pre-processing, execute a few variations of model training in parallel and aggregate multiple results, hyperparameter optimization, distributed training and model serving.” [Oliveira et al. 2024]
Related terms: inference, deployment, TensorFlow Serving