Arize AI
Search…
1. Setting Up Your Account
Familiarize yourself with Arize concepts and names

Overview

Before setting up your account, let's go over some important concepts and terms you will see in the Arize platform. Your Arize account can consist of multiple organizations and spaces. Within each organization you may have several spaces which, in turn, may have multiple projects to organize your groups of models.

Definitions and Examples

  • Account: You are a member of one Arize account. An account consists of one or more organizations.
    • Example: Ridesharing Company A
  • Organizations: Organizations represent a single business unit and help you silo work across different areas of your business. Within your account, you can be a member of multiple Organizations. An Organization may consist of one or more spaces.
    • Example: Risk Organization, Research Organization, etc.
  • Spaces: Spaces represent an environment for groups of models. They help promote collaboration across models within an organization and can be used to experiment in a safe environment without polluting the organization. You can be a member of multiple spaces, across multiple organizations within your account.
    • Spaces can either be public or private.
    • Public Spaces: Visible to all members (regardless of role) of the parent organization.
    • Private spaces: Only visible to explicitly invited members of the space.
    • Example: Payment Fraud Space
  • Project: Projects are optional. A project is a way to group or organize multiple models into a folder within a space. When leveraging projects, you can view performance across many models.
    • Examples: Marketing Project, Pricing Project, Forecasting Project
  • Model: This is the most important entity in Arize. Each model has its own set of monitors and configurations set up. Any unique model you want to view monitors on should be at the model level. These monitors will apply for all model versions.
    • Example: Pricing Model for San Francisco
      • San Francisco has its own model because I want to get alerted if SF's model performance goes down
  • Model Version: Each model can have many versions. A version is sent in with each inference. Learn more here.
    • Example: Pricing Model for San Francisco v100
  • Model Environment: Arize supports 3 environments. Learn more here.
  • Model Schema: Model schema is not static. As new features are added, Arize schema updates for that model. Learn more here.
Questions? Email us at [email protected] or Slack us in the #arize-support channel
Copy link
Outline
Overview
Definitions and Examples