DataRobot announced new enhancements to its enterprise AI platform. The new features are designed to make it easier for every user, from advanced data scientists to non-technical, front-line decision makers, to derive value from AI.
“Our research has shown that early adopters of AI report improved customer experience, accelerated rates of innovation, higher competitiveness, higher margins, and better employee experience, yet despite the benefits, many struggle with deploying AI. To truly get the most value from AI, it needs to be trustworthy, scalable, and easy to manage,” said Ritu Jyoti, Program Vice President, WW AI and Automation Research Practice at IDC.
“DataRobot’s new capabilities will open up the power of AI to even more personas, including advanced data scientists, and empower organizations to work on data science initiatives in a highly collaborative way, ultimately improving front-line decision making through higher quality AI powered applications.”
Specifically, the new enhancements include:
Created with the most advanced data scientists in mind, the Composable ML feature allows AutoML users to clone, edit, and reconfigure DataRobot’s blueprints to fit the specific needs of their use case.
Composable ML opens the hood to DataRobot’s world class automation and unlocks the blueprints provided in its repository to granular levels of configuration.
Users can also integrate their own custom training code to create entirely new models that instantly work with DataRobot’s explainability tools and have a clear path to production via DataRobot MLOps.
To ensure every model put into production remains accurate and viable, DataRobot has extended the power of its MLOps product with Continuous AI: a feature that will allow users to set up multiple retraining policies on their production models.
With Continuous AI, users can schedule their models to be automatically retrained on a regular basis or when an event like data drift occurs. The feature will also leverage DataRobot’s AutoML capabilities to automatically create new challengers ensuring the best, most accurate model is always available for use.
Continuous AI operates within the existing MLOps governance framework, ensuring no production models are updated or replaced without passing through a gated approval process.
No Code AI App Builder
The No Code AI App Builder allows users to quickly turn any model into an AI application, without requiring any coding. Drag and drop widgets, data visualizations, and pre-built templates enable the creation and deployment of powerful new AI apps in a matter of minutes.
The No Code AI App Builder makes it much easier for business users and front-line decision makers to leverage the predictions generated by their models to make more informed, AI-backed business decisions.
Bias and Fairness Production Monitoring
Building on its Bias & Fairness Testing feature, DataRobot created Bias & Fairness Production Monitoring, which proactively monitors production models for bias.
With this addition, DataRobot’s platform enables end-to-end bias testing and monitoring, ensuring every model that is created and put into production is trusted and fair. The platform will alert users whenever bias is detected and provide guidance on the factors that cause bias to mitigate recurrence.
A new tool to evaluate existing AI models and generate an automatic scorecard grading them across four critical areas: data quality, robustness, accuracy, and fairness.
For each grade, detailed explanations are also provided, enabling customers to understand if their models are best-in-class and ready for production.
“We’ve always seen AI as a team sport, and to truly democratize its capabilities we need to serve the most technical and the most urgent business audiences,” said Nenshad Bardoliwalla, SVP of Product at DataRobot. We’re opening up our platform to enable advanced data scientists to explore their own custom-built code, while simultaneously delivering no-code solutions to empower non-technical business users with AI at their fingertips. These investments allow us to truly tackle all the personas necessary to make AI pervasive.”