IBM Developer Blog

Follow the latest happenings with IBM Developer and stay in the know.

New suite of integrated developer productivity tools makes scaling AI for the enterprise more attainable than ever


The most common reason an AI model fails is not because the model itself is flawed, but from lack of proper data preparation and organization. AI is not magic, but it does require a thoughtful, well-planned information architecture and approach.

With IBM Cloud Pak for Data, developers can now build and deploy end-to-end, real-time data science use cases in a cohesive environment that delivers scalability and security.

It enables:

  • Self service — Easily spin up scalable instances of supported services without needing to ask a sysadmin for help.
  • Collaboration — Work together in an integrated environment with projects and other shared assets.
  • Asset management — Manage shared assets and track provenance over time.

Best of all, it runs natively on the Red Hat OpenShift Container platform — which provides:

  • The ability to move the environment and applications anywhere that OpenShift is supported without re-writing code
  • Hybrid- and multi-cloud support
  • Easy integration with the large and growing ecosystem of other components that run on OpenShift
  • Easy management of underlying, shared Kubernetes resources to allow for better scalability and security, and more effective use of the underlying compute and storage

In line with our Watson Anywhere vision, the latest release of Cloud Pak for Data makes the world-class Watson services available in private, hybrid, and multi-cloud environments.

Cloud Pak for Data also has a range of new features that can help jump-start your use case development. The latest version includes these advanced Watson features:

  • Watson Assistant to enable intelligent conversational interfaces
  • APIs for speech-to-text and natural language understanding
  • Watson Discovery for gaining insights from collections of unstructured documents (check out the demo)
  • AutoAI, which uses AI to automate many data science tasks, such as data preparation, model development, feature engineering, and hyper-parameter optimization (check out the article “AutoAI: Humans and machines better together“)
  • Watson OpenScale, which checks your models for bias and alerts you in case of drift (check out the “Getting started with IBM Watson OpenScale” tutorial)

It also includes the following new open source packages for data and AI:

  • Take advantage of the latest versions of data science and machine learning frameworks like TensorFlow, PyTorch, scikit-learn, XGBoost, and pandas
  • Analyze data, train models, and write code in a full JupyterLab environment
  • Leverage IBM’s production-hardened Spark on Kubernetes for scalable ETL, SQL, and machine learning workloads
  • Spin up and leverage data stores like Postgres and MongoDB

In addition, Cloud Pak for Data offers:

  • The ability to build advanced, high-performance stream processing applications using a drag-and-drop interface
  • World-class developer tooling from our partners like Lightbend and RStudio

Check out our “Getting Started with Cloud Pak for Data” learning path to find out how to virtualize, visualize, and transform your data, build and deploy your own machine learning models, and leverage them in your applications with continuous monitoring to ensure quality of outcome.

Right now, you can test drive IBM Cloud Pak for Data free for 7 days — learn more here.

Find all the details on these developer resources and more in the Cloud Pak for Data IBM Developer Hub.

Willie Tejada