Projects & Environments¶
A project is a scoped container for AI service development, pipelines, and related artifacts. Projects must be created inside an existing workspace, inheriting its region, environment tag, and resource boundaries.
Creating a Project¶
- In Infra Hub, click Projects →
+ Project - Enter project name and description
- Select the parent workspace
- Assign users and roles (Platform Admin, AI Engineer, QA, Architect, Auditor)
- Select environment(s) to link to the project
- Submit
Environment Types¶
| Environment | Purpose | Characteristics |
|---|---|---|
| Development (Dev) | Feature development, experiments, initial debugging | Smallest compute footprint. Rapid, unchecked deployments permitted. |
| Quality Assurance (QA) | Systematic testing, validation, integration checks | Medium instances. Functional, regression, and integration test suites. |
| Staging | Near-production replica, load testing, user acceptance | Matches production architecture. Same container images as Prod but isolated from live traffic. |
| Production (Prod) | Live environment serving real users | Largest compute allocation. Full disaster-recovery and scaling policies. Only approved service versions deployed. |
Info
The environment within a workspace is shared across all projects in that workspace. If a workspace has a Medium environment, all projects under it share that compute pool.
Switching Projects¶
Use the project switcher (top-left dropdown) → select Config view → click the switch icon → select the desired project and environment → Submit.