Multivocal Literature Review & Interviews: Continuous End-to-End Lifecycle Management Pipeline for Artificial Intelligence

During a multivocal literature review 151 relevant formal and informal sources were extracted. Additional information was collected from nine semi-structured interviews.

The extracted dataset provides the foundation for the presentation and comparison of different terminologies for DevOps and CI/CD for AI, MLOps, (end-to-end) lifecycle management, and CD4ML. Furthermore, the dataset comprises potential triggers for reiterating the pipeline and consolidated tasks necessary in a continuous end-to-end lifecycle pipeline categorized into four stages: Data, Model, Dev and Ops. Additionally, the dataset provides information regarding challenges of the lifecycle pipelines for AI.