MLOps
Technology & Computing abbreviation · Artificial Intelligence
| Stands for | Machine learning operations |
|---|---|
| Categories | Technology & Computing |
MLOps stands for Machine learning operations, an artificial intelligence abbreviation.
The practices and tools for deploying, monitoring, versioning and retraining machine learning models reliably in production, applying DevOps ideas to the machine learning lifecycle.
Example: Our MLOps pipeline retrains the model weekly and blocks deployment if accuracy drops.
Usage notes
The term spread around 2018 to 2020, partly following Google's 2015 paper "Hidden Technical Debt in Machine Learning Systems". Common tools include MLflow, Kubeflow, SageMaker and Weights & Biases. Monitoring for data drift is a core concern, since models degrade as real-world data changes. LLMOps applies similar ideas to language-model applications.
Related abbreviations
More artificial intelligence abbreviations
- MAE – Mean absolute error
- mAP – Mean average precision
- MCP – Model Context Protocol
- MDP – Markov decision process
- ML – Machine learning
- MLE – Maximum likelihood estimation
- MLP – Multilayer perceptron
- MMLU – Massive Multitask Language Understanding
- MoE – Mixture of experts
- MSE – Mean squared error
- MT – Machine translation
- NER – Named-entity recognition