Analysts, data engineers, and data scientists have always been core roles that contribute to advanced analytics projects. But in order for an organization to scale AI initiatives with a more systematic approach, the development, operationalization, and oversight of AI projects must also include contributors from different parts of the organization, including IT operators, project managers, risk managers, and subject matter experts (SMEs). Dataiku 10 has exciting new capabilities that accelerate time to value and empower people across diverse functions to engage in data projects and responsibly deliver and manage AI applications.
Deploy and Maintain More Models in Production
With Dataiku 10, data scientists and IT operators have additional tools and more flexibility to deploy, monitor, and manage machine learning (ML) models at scale. This release features built-in drift monitoring and alerts, as well as a model evaluation store to capture and visualize performance metrics over time. Once a model is developed or retrained, a model comparison tool aids with Champion/Challenger analysis to help operators determine the best course of action and continuously improve outcomes.
Furthermore, in addition to models developed natively in Dataiku, models developed externally in MLFlow can now also be deployed, managed, and governed in Dataiku. Risk and operations teams will appreciate Dataiku’s robust MLOps framework that orchestrates model delivery activities in a central location and reduces the burdens of manual model oversight and routine maintenance.
Safely Scale Initiatives With Oversight
Dataiku 10 also introduces new capabilities for AI Governance, project workflows, reviews and sign-off, and portfolio oversight for enterprise customers. Key business and project stakeholders can first use a standardized framework to assess potential project value and risks, both to compare initiatives for investment and determine appropriate oversight levels.
Customizable project plans with clear steps and a sign-off feature for model deployment enable analytics teams to explore, build, test, and deploy projects with the appropriate reviews and approvals. Finally, the model registry provides a centralized way to see all models (whether developed in Dataiku or externally) in one place, versioned and with performance metrics and project summaries. The result of this multifaceted governance strategy is increased transparency, trust, and confidence in both the analytical process and the system’s outputs.
Deliver Value Faster With Business Solutions and Accelerators
Organizations can achieve faster speed to value by leveraging Dataiku Industry Solutions, off-the-shelf starter projects for vertical use cases that teams can adapt and apply for their own purposes. Enhanced visual statistics, geospatial analytics, and new visualization capabilities make it even easier for designers to explore and enrich data and distribute interactive insights to business stakeholders. Dedicated workspaces provide AI consumers with a single point of access to a variety of analytic outputs, facilitating adoption of Everyday AI. Finally, data scientists and SMEs alike will love the additions to interactive scoring, taking what-if analysis to the next level by enabling teams to prescribe the specific changes to inputs needed in order to influence a different outcome.
Dataiku is the platform for Everyday AI, systemizing the use of data for exceptional business results. With Dataiku 10, organizations can involve more users across a diverse set of roles (i.e., IT operations, domain experts, risk managers) to ultimately create, deliver, and manage more high-value projects at scale. Ready to dive into the details on Dataiku 10? Check out the full list of resources on what's new.