Momentum MLOps provides CI/CD for ML pipeline so that data engineers and scientists can deploy models from their development environment into production.

Data Scientists can continue using tools and programming languages of their choice to develop models and be able to efficiently deploy, manage, monitor, and scale their models in production. Momentum MLOps also provides web based user interface to manage the deployment, access control, prediction, and monitoring.

Model version control, A/B testing, data drift detection, model drift detection, team collaboration, and model governance are some of the key features that streamline the overall enterprise ML process. 

Release Date: January 10, 2022, version 2.0

1. Web based user interface to sign up
2. Access control and roll based access
3. Model registry to provide a view of all deployed models and corresponding metadata
4. Model deployment via web UI
5. Model deployment via Curl command
6. Model deployment via Restful API
7. Model deployment via Python and Java
8. Get service to get the model artifact and metadata
9. Web interface to annotate a deployed model
10. Upload model training/ reference dataset via browser upload
11. Delete a model from the web interface
12. Display a set of graphs to display model operations
    1. Number of model consumers
    2. Number of prediction requests received
    3. Number of predictions served
    4. Median response time per prediction request
    5. Data error rate
    6. System error rate
    7. Trend of daily data error rates
13. Data drift detection
    1. Display predictors or fields having any drift compared to the reference data
    2. Display the probability distribution and P-values of each predictors/fields
14. Users and role management
    1. Edit user profile
    2. Change password
    3. Add new users
    4. Add roles and permissions to a user
15. Role based access
    1. Restrict users from accessing models and metadata if the user does not have privilege to that model
    2. Allow users to share their own models with other users within the same organization
    3. Restrict users from accessing features based on roles and privileges
16. APIs
    1. Allow users to display the API key
    2. Allow users to change their API key