Top features from Dataiku 11 that will help accelerate enterprise AI projects

Top features from Dataiku 11 that will help accelerate enterprise AI projects ...

Dataiku, a New York-based company that provides a centralised solution for the design, deployment, and execution of enterprise artificial intelligence applications, has released version 11 of its unified data and AI platform. The update, which is expected to be generally available in July, focuses on delivering on everyday AI''s promise and provides new capabilities to help data experts handle larger, multidisciplinary tasks.

Expert data scientists, data engineers, and ML [machine learning] engineers are among today''s most sought-after tasks. Most of their time, talented data scientists spend most of their time developing low-value logistics such as establishing and maintaining environments, preparing data, and putting projects into production. These are key things they can do to help businesses become more successful and shape a culture of AI to transform industries, according to Clement Stenac, the CEO of Dataiku.

The following is a breakdown of the key capabilities.

Code Studios with experiment tracking

In Dataiku 11, AI developers are provided with a fully managed, isolated coding environment, which allows them to work with their own IDE or web application stack. This way, AI developers can demonstrate how comfortable their organization is when it comes to analytics centralization and governance (if any). Previously, anything like this would have meant going for a custom setup, with increased costs and complexity.

Transform 2022

In person, please join us at the leading event on enterprise business and technology decisions on July 19 and in the summer 20-28.

The solution also comes with a learning curve, which allows developers to have a central interface to store and compare all programmatic models created using the MLFlow framework.

Seamless computer vision development

Dataiku 11 includes a built-in data labeling framework and a visual ML interface to simplify the resource-intensive task of developing computer vision models.

The former, according to the company, annotates data in large amounts, a task often handled through third-party websites like Besides, the latter gives a end-to-end, visual path for common computer vision tasks, allowing both advanced and novice data scientists to deal with complex object detection and image classification applications, from data preparation to developing and deploying the models.

Time-series forecasting

Business users, particularly those with limited technical knowledge, often find it difficult to analyze historical data and create robust business forecast models for decision-making. Dataiku 11 offers built-in capabilities that provide no-code visual interfaces and assist teams analyze temporal data and develop, evaluate and deploy time-series forecasting methods.

Feature Store

The latest release also includes a Feature Store that provides new object-sharing capabilities to improve organization-wide collaboration and accelerate the entire process of model development. According to the company, the capability will give data teams a dedicated area to access or share reference datasets containingcurated AI features. This will prevent developers from using redundant data assets for ML projects, and prevent inefficiencies and inconsistenencies.

Outcome Optimization

Teams may use a manual trial and error (what if) method to provide actionable insights to their business stakeholders that might assist them in achieving the most desired results.

Outcome Optimization will be implemented as part of Dataiku 11, and it will take into account user requirements and select the optimal set of input values to achieve desired results. For example, it might specify what changes a manufacturer might make to factory conditions in order to obtain maximum production yield or what adjustments to a bank consumers financial profile would result in lower risk of loan defaults.

Other capabilities

Among other things, the company has expanded its surveillance and control of model development and deployment. This includes a powerful software to generate flow documents and a central registry that records snapshots of all data pipelines and project artifacts for review and sign-off before production. The company will also provide model stress tests that will investigate model behavior before the deployment.

You may also like: