Updates from Dataiku Enhance Deployment, Visualizations, and More

Dataiku, the platform for Everyday AI, is unveiling a myriad of updates and new integrations that build off of the company’s previously announced generative AI (GenAI) capabilities in October. The new features—including new deployment options, Databricks integrations, visualizations, and more—continue to enhance the Dataiku experience, according to the company. 

The first of Dataiku’s updates is a “deploy anywhere” capability which enables users to deploy an API service developed within Dataiku to other production environments outside of Dataiku API nodes. 

These new destinations for deployment—including AWS SageMaker, Azure ML, and Google Vortex—continue to position Dataiku as an open, infrastructure-agnostic solution for monitoring, governing, and democratizing access to all of a user’s models, according to the company. 

Expanding the integration with Databricks, Dataiku users are now able to surface models from Databricks as “external models” in Dataiku, as well as import MLflow models directly from a Databricks model registry or Unity Catalog. This expansion drives greater explainability and visibility via interactive model reports, performance comparisons, and scoring capabilities. 

Dataiku is also introducing improvements to its dashboards, including new chart types, more business context, and better interactivity. New sankey diagrams enhance resource flow and process path visualizations, while measure-based reference lines applicable to charts and cross-filtering in dashboards enhance business understanding. 

Other updates to Dataiku include:

• Model overrides option for more control over model responses
• Exporting statistical tests as recipes for greater operationalization and automation
• Importing pre-labels for text labeling/validation and an upgraded NER recipe
• New tutorials in the Dataiku Academy and Developers Guide
• Pre-built business solutions for credit risk stress testing and predictive maintenance use cases