Skip to main content

What is H2O AI Cloud?

H2O AI Cloud (HAIC) is an end-to-end AI platform designed to help create, deploy, monitor, and share data models or AI applications for virtually any use case. It is deployed on Kubernetes and can be run on most cloud or on-premise infrastructure. Depending on how much control you want to have over your HAIC deployment, you can choose between the two available deployment options.

The platform encompasses a few core components and products of H2O.ai that are installed within the environment by default, and a few extra, paid components that can be included upon request.

Base components that are enabled by default:

  • AI App Store
  • H2O AI Engine Manager
  • H2O Driverless AI
  • H2O-3
  • Wave SDK
  • H2O MLOps
  • H2O Drive (optional)
  • H2O LLM Studio (optional, open-source version)
  • H2O Admin Analytics (optional)
  • H2O Orchestrator
  • H2O Platform Services
    • AuthZ
    • H2O Drive backend service
    • Cloud Discovery
    • Token Server
    • Secure Store
    • Model Hub
    • Telemetry

Apart from these core components, there are also a few optional components that you can install as part of your H2O AI Cloud environment for a paid subscription.

Paid subscription components:

  • H2O Hydrogen Torch
  • H2O Label Genie
  • H2O eScorer
  • H2O Document AI
  • H2O Feature Store
  • Enterprise h2oGPTe
  • H2O LLM Data Studio
  • H2O LLM Eval Studio

Let's begin by understanding what you can do using HAIC and how the preceding components fit into the big picture to run seamlessly together as the single platform that is HAIC.

Navigate to H2O AI Cloud and log in.

H2O AI Cloud homepage​

Once you are logged in, the H2O AI Cloud homepage is the first page that you see.

H2O AI homepage

The dropdown on the top-left corner of the screen lets you choose the Workspace that you want to work with.

A Workspace is a mechanism for resource-sharing and collaboration within the H2O AI Cloud platform. It hosts specific app instances that you can share with members of your team so that multiple people can collaborate on a single app instance.

Workspaces

To create a new Workspace or add new members to your Workspace, click the dropdown arrow and click Manage Workspaces.

Once you have selected your Workspace, you can navigate through the rest of the homepage. The homepage consists of different widgets to allow the users to access the products of H2O AI Cloud quickly and efficiently.

  • Top widget - Start here : This will be different for each environment, and the admin can change the title text and the app details pages that are linked from here.

  • Center left widget - My app instances : This will show your users' latest app instances. From here, you can visit or pause a running instance and resume a paused instance. Additionally, you can pause or resume all the instances at once by clicking on ... on the top right corner of the widget and selecting Resume all instances or Pause all instances. To view all the app instances on the My instances tab, click View all.

  • Center center widget - My AI engines : This will show your users' latest AI engine instances. From here, you can visit or pause a running instance and resume a paused instance. Additionally, you can pause or resume all the AI engines at once by clicking on ... on the top right corner of the widget and selecting Resume all engines or Pause all engines. To view all the AI Engines on the My AI engines tab, click View all.

  • Center right widget - My pinned apps : You can pin any Apps that you want to be able to access quickly by clicking on the pin icon found below each app listed in the App Store page. All the pinned apps will be shown here.

  • Bottom left widget - Learn to make apps : This will be different for each environment, and the admin can change the title text and the app details pages that are linked from here.

  • Bottom right widget - Latest uploaded apps : This will show the latest apps that are uploaded to the AI App Store. To view a specific app details page, click View.

AI App Store​

Once you are logged in, click App Store on the side panel to visit AI App Store.

AI App Store Screen

The AI App Store is a front-end platform and deployment system used to rapidly prototype, deploy, and browse AI applications that are designed for fulfilling different use cases. Your organization can have its own internal App Store to host the applications that are most important to your company. On this App Store, you can deploy or browse AI applications built by your company's engineers, as well as include applications built by the makers at H2O.

The categories panel that you see when you hover over App Store lists out the various categories of applications that are available on your App Store, or you can simply click App Store and use the search bar at the top to search for a particular app.

Build models and experiments​

H2O AI Cloud enables you to create new AI models and run experiments on your data using automated machine learning capabilities such as feature engineering, validation, selection, and tuning for your model. Since the key machine learning tasks are automated, this helps data scientists to work more efficiently on their projects.

The automated machine learning capabilities are powered by Driverless AI and H2O-3, which are in turn managed by H2O AI Engine Manager. AI Engine Manager is responsible for managing the lifecycle of an AI Engine.

Click AI ENGINES to create a new AI Engine or author a new Notebook.

My AI Engines

App development​

H2O AI Cloud provides an open-source, Python app development framework called H2O Wave, which you can use to build business-friendly AI applications using the models that were built and productionalized using the other components of the HAIC platform. The Wave SDK allows developers to build easily using Python libraries without having to worry much about the front-end JavaScript or CSS for the interactive UI components.

Once you have built your application, you can navigate to Apps to import it to H2O AI Cloud.

My Apps

To view and access your active app instances, click App Instances. An app instance will be paused automatically if it has not been used in approximately three hours.

Experiment with data​

H2O AI Cloud provides a collaborative development environment where you can author and access Jupyter notebooks from a single lab environment. You can use Notebook Lab to explore data, write code, and experiment with your data.

Click Notebook Lab to work with Jupyter notebooks.

Model validation and monitoring​

The data models that are created are productionalized by H2O MLOps, where the model validation and monitoring of its performance takes place so that it can then be consumed by an AI application.

Click Models to navigate to H2O MLOps.

H2O MLOps

Use AI for documents​

H2O AI Cloud also offers a few components on paid subscription that you can use to process, interact, and gain insights from your documents and data using generative AI:

  • Enterprise h2oGPTe - a secure, enterprise-grade generative AI platform that provides organizations with a complete solution for implementing AI capabilities across their operations. It can be used as an LLM assistant to query collections of documents, extract insights from them, and interact with them.

  • Document AI - an AI engine used for classifying, extracting, analyzing, refining, and labeling data from unstructured documents.

  • Eval Studio - a modular and extensible studio for RAG and LLM evaluation, used to evaluate and test your models.

Feedback