Paving the path to enterprise AI

It also welcomes deep learning to its supported libraries, improves AutoML, and introduces new comprehensive documentation features that allow for global, distributed data teams to build powerful applications more efficiently.

Florian Douetteau, CEO of Dataiku, stated: “In Dataiku 5 we are unlocking the ability for global teams to build even more powerful AI-driven services. We envision the future of AI application development as a global network of people within an organisation, all of different backgrounds and skill levels, contributing to game changing business applications. We’ve taken a major step towards that vision in our latest release.”

Dataiku 5 enables scalable Enterprise AI across an organisation by introducing a host of new features, namely:

Full containerisation with Docker and Kubernetes

‘Containerised’ environments are self-contained entities that can be moved to various servers and allow organisations to load models and serve requests in a few clicks. Dataiku 5 takes Docker and the power of ‘containerisation’ to in-memory processing of Python and R recipes, as well as in-memory model training and scoring.

To run in-memory jobs, Dataiku can now create a Docker image containing the code and required libraries or packages, and automatically deploy it to a Kubernetes cluster for computation elasticity, allowing for more scaling and easier isolation of resources.

Maritime Containers were a revolution in shipping that shaped the global economy as we know it today because they enabled a global network of transportation for goods. Similarly, Virtual Containers are changing the way we develop and deploy applications in the Enterprise. With maritime containers, the week point is always the first and last mile: how can you fill container and ultimately deliver the goods to the end customers.

Similarly, in order filling containers with AI capabilities is today requires to master and connect complex technologies. Delivering machine learning in containers with a click is the revolution being created with Dataiku 5. 

Visual deep learning with TensorFlow and Keras

Dataiku 5 welcomes Keras, the neural network library written in Python, to its ever-growing family of supported open-source libraries and technologies. Users can now define the architecture of their deep learning models directly from the Machine Learning GUI. Plus, due to the availability of code samples and Tensorboard, even the members of the team who aren’t familiar with deep learning architectures can now participate in building and monitoring these more complex models and AI applications.

Comprehensive documentation with group discussions and Wikis

As organisations are creating more AI-enabled services, the lack of documentation creates risks and rising maintenance costs. Encouraging strong documentation in AI projects promotes sustainable reuse of knowledge from one project to another and to allow teams to develop the skills critically needed to be successful. Users can now easily share and find knowledge with ‘Dataiku Discussions’. In Dataiku 5, anyone can create, subscribe, comment, interact with teammates or subject matter experts on their chosen topics of interest. In addition, organisations can now create Wiki spaces within Dataiku for collaborative creation and editing of documentation.

Enterprise AI is on track to change business processes and efficiencies in the near future and Dataiku 5 is the launching point for data teams to begin creating powerful applications all on one single collaborative platform. Whether all in a single office or distributed across difference continents, the powerful collaboration features inherent in the Dataiku platform make it ideal for organisations looking to scale their data operations. 

Currently, hundreds of companies use Dataiku on a daily basis to enable their teams to build, deploy, and monitor predictive data flows and build enterprise AI applications that solve industry-wide problems like fraud, churn, supply chain optimisation, predictive maintenance, and much more.