Cloudera Delivers Open Standards Based MLOps Empowering Enterprises to Industrialize AIMay 6, 2020
Cloudera, the enterprise data cloud company, today announced an expanded set of production machine learning capabilities for MLOps is now available in Cloudera Machine Learning (CML). Organizations can manage and secure the ML lifecycle for production machine learning with CML’s new MLOps features and Cloudera SDX for models. Data scientists, machine learning engineers, and operators can collaborate in a single unified solution, drastically reducing time to value and minimizing business risk for production machine learning models.
“Companies past the piloting phase of machine learning adoption are looking to scale deployments in production to hundreds or even thousands of ML models across their entire business,” said Andrew Brust, Founder and CEO of Blue Badge Insights, an independent advisory firm. “Managing, monitoring and governing models at this scale can’t be a bespoke process. With a true ML operations platform, companies can make AI a mission-critical component of their digitally transformed business.”
The release of Cloudera Machine Learning with new MLOps features and Cloudera SDX for models provides a fundamental set of model and lifecycle management capabilities to enable the repeatable, transparent, and governed approaches necessary for scaling model deployments and ML use cases.
- Unique model cataloging and lineage capabilities allow visibility into the entire ML lifecycle to eliminate silos and blind spots for full lifecycle transparency, explainability and accountability.
- Full end-to-end machine learning lifecycle management that includes everything required to securely deploy machine learning models to production, ensure accuracy, and scale use cases.
- A first-class model monitoring service designed to track and monitor both technical aspects and accuracy of predictions in a repeatable, secure, and scalable way.
- Built on a 100% open source standard and fully integrated with Cloudera Data Platform, enabling customers to integrate into existing and future tooling while not being locked into a single vendor.
“Cloudera has been working across our industry and with some of our largest customers and partners to build open standards for machine learning metadata,” said Arun Murthy, chief product officer, Cloudera. “We have implemented those standards as part of Cloudera Machine Learning to deliver everything enterprises need for deploying and sustaining machine learning models in production at scale. With first-class model deployment, security, governance, and monitoring, this is the first end-to-end ML solution for full-lifecycle management from data to ML driven business impact across hybrid and multi-cloud.”
The expanded set of production machine learning capabilities available in Cloudera Machine Learning (CML) include:
- New MLOps features for monitoring the functional and business performance of machine learning models:
- Detect model performance and drift over time with native storage and access to custom and arbitrary model metrics.
- Measure and track individual prediction accuracy, ensuring models are compliant and performing optimally.
- Cloudera SDX for models extends SDX governance capabilities to now support models:
- Track, manage, and understand large numbers of ML models deployed across the enterprise with model cataloging, full lifecycle lineage, and custom metadata in Apache Atlas.
- View the lineage of data tied to the models built and deployed in a single system to help manage and govern the ML lifecycle.
- Increased Model security for Model REST endpoints, which allows models to be served in a CML production environment without compromising security.
Availability and Pricing
Cloudera Machine Learning with new MLOps features and Cloudera SDX for models is available on CDP for both Microsoft Azure and Amazon Web Services as an integral part of the Cloudera Machine Learning platform. CML is charged by the hour and starts at $0.68/hr per instance. Detailed CDP and CML pricing information can be found here.