Defining MLOps

<aside> 💡

An all-inclusive, holistic approach to managing ML systems

</aside>

The set of practices, processes, and technologies for managing data, code, and models to improve performance stability and long-term efficiency in ML systems.

Responsibilities of MLOps Components

<aside> 💡

Key Functions of DataOps, DevOps, and ModelOps in MLOps

</aside>

DataOps

• Optimized data processing
• Centralized data discovery, management, and governance
• Ensured data quality
• Traceable data lineage and monitoring

Devops

• Machine learning is code
• Continuous integration and continuous deployment (CI/CD)
• Version control via GIT
• Production-grade workflows

ModelOps

• Move beyond models as objects
• Treating model code as software
• Treating models as data
• Manage the model lifecycle