for a paper at university I am currently evaluating Dataiku as a MLOps platform using a questionnaire. However, I could not find an answer to some questions for Dataiku so far. Therefore, I would be very happy if you know an answer to the following questions:
- Can the UI of the dashboard be configured?
- Are there dedicated and different pipelines for CI/CD and modeling?
- Regarding scalability and high availability: What machine sizes can be used? And how fast can machines be shut down in order to save costs?
- Is there an autoscaling functionality for endpoints?
- Did I understand correctly that in Dataiku, Git enables full reproducibility by tracking (tracking of model hyperparameters, training pipeline configuration parameters, model training data and version, source code and container versions used in pipelines)?
Any feedback would help me a lot. Thanks in advance!