Skip to main content
Choose New project to answer six questions about your task, complexity, simultaneous users, answer speed, input size, and infrastructure. The platform recommends a model and GPU configuration based on those requirements. Review the recommendation and alternatives, name your project, and choose Create project. Use Add details to describe the use case or indicate that you already have a fine-tuned model. Follow the quickstart for the full sequence.

​
Read the recommendation

The suggested model, precision, GPU count, and performance figures are planning estimates. They help define a setup to validate with representative requests; they are not a deployed model, reserved capacity, or a performance guarantee.

​
What to include

  • Task: for example, extract policy numbers and loss dates from claim notices.
  • Model: build a new specialized model, or bring one already trained elsewhere.
  • Data: use a starter or prepare your own supported records.
  • Locations: choose preparation/training and serving destinations separately.
  • Success criteria: define correct outputs, missing-value behavior, and acceptance checks.
  • Budget: record the intended experiment budget. This is a planning value, not a compute reservation or enforced dollar cap.
The current setup uses structured questions and the platform’s sizing planner. These answers and the use-case description are not sent to an external AI provider. Use Data for supported source records and confirm processing requirements before adding confidential information.

​
Review and resume

Open Overview after creating the project and follow Continue setup when shown. Review the task, data, and processing locations, then save the project plan. Project records belong to your team. Your saved plan is the brief for model setup; inference becomes available after a model is connected to the project. If someone else edits the same version, the platform rejects the stale save rather than overwriting their changes. Reload the latest version before continuing.

​
Project console

The project list shows the registered model, runtime status, and GPU telemetry when the connected runtime reports it. Missing telemetry is not zero utilization. Admins and developers can edit projects. Members can read them and use a connected model, but cannot edit setup or start and stop demos.