Tools for automating ML Workflows in Azure for faster deployment of ML in production
Azure Machine Learning service’s python based SDK, provides data scientists and developers to integrate different sub-processes of an ML workflow and deploy as a solution to the user. Hugely frugal in terms of computational usage during CI/CD activities.
A very powerful open-source platform used to create, schedule, and monitor ML workflows at scale. With workflows defined as Directed Acyclic Graphs (DAG), dependency management of different tasks in the pipeline is ensured.
A Kubernetes tool specifically meant for creating and managing ML pipelines. Allows specifying DAGs, where each step of DAG is a Kubernetes pod. Python language is used for interaction with Kubeflow.
A tool that is part of Azure Databricks, it has predefined patterns for tracking experiments and deploying models. MLFlow is a Python library that can be imported to the existing ML code and a CLI tool can be used to train and deploy ML models written in scikit-learn to AWS SageMaker or Azure Machine Learning Service.
Tools for monitoring ML models on Azure for quick monitoring of ML in production
A go-to tool that offers monitoring capabilities for all ML stakeholders at all value-creation points of ML lifecycle, and that works on all cloud environments. Provides a neat-and-clean, and modular observability experience to keep the ML model delivering the right value to the user.
In addition to the creation of ML workflows, Azure Machine Learning SDK also can be used for defining drift-monitoring tasks between datasets using Python classes like DataDriftDetector and AlertConfiguration.
Anodot’s MLWatcher tool is used for collecting and monitoring metrics from ML models in production. This open source Python agent is free to use, where a BI tools can be connected to visualize the results
A tool that helps to monitor ML Model performance during training, track predictions, and capture drift. Alerts and trigger actions can be set for notification when model training is completed or performance begins to drift.
Gain insights from production ML input/predictions data, logs and application data to continually improve your model performance
Please feel free to schedule a demo for data quality assessment with us or try Qualdo now using one of the team editions below.