The SAFE-6G MLOps framework streamlines the full lifecycle of machine learning (ML) models—from development to deployment—leveraging modern cloud-native technologies. It automates workflows, ensures reproducibility, manages versioning of models and data, and incorporates differential privacy to secure sensitive data. The framework utilizes tools like GitLab CI/CD, Kubeflow, and Kubernetes for continuous integration and deployment, model orchestration, and scalable serving. Key components include a pipeline development environment, orchestration platform using DAGs, and automated model creation with differential privacy. Models are stored and deployed as containerized microservices, ensuring efficient, secure, and dynamic ML operations across edge-cloud environments.
Transcript.

Find it here: https://www.youtube.com/watch?v=-mfUbtJZvwU&ab_channel=SAFE-6Gproject