Creating Your First Project

Step-by-step guide to create a project and operationally prepare it

This guide walks through a practical setup flow for your first Bud AI Foundry project.

Before You Start

Make sure you have permission to create and manage projects in your workspace.

1. Create the Project Record

1. Open Projects.
2. Click + Project.
3. Fill in:
   • Name: Fraud Detection Assistant
   • Description: Project for fraud triage model APIs
   • Tags: risk, production, llm
4. Click Next to create project.
5. Proceed to the Invite members step.

2. Add Team Members

1. Open the new project.
2. Click Share/Invite.
3. Add users and assign view/manage scopes.
4. Confirm invited members appear in project access controls.

3. Deploy a Model in the Project

1. Go to the Deployments tab.
2. Create a deployment and select:
   • Model source (for example Hugging Face or cloud provider)
   • Target cluster
   • Deployment name
3. Save and wait for deployment status to become healthy/ready.

4. Configure Deployment Settings

Open deployment Settings and configure:

• Rate limits to protect capacity
• Retry/fallback rules for resiliency
• Autoscaling minimum/maximum replicas and trigger metrics

5. Add Route Logic

1. Open the Routes tab.
2. Create a route and select strategy (for example simple shuffle or least busy).
3. Attach one or more eligible deployments.
4. Save and validate route status.

6. Validate Analytics and Observability

1. Open Analytics to inspect baseline latency/throughput.
2. Open Observability to verify runtime event visibility.
3. Confirm traffic and error patterns are visible before broad rollout.

7. Apply Guardrails

1. Open Guardrails.
2. Attach required safety or compliance policies.
3. Verify policies are active and aligned with project requirements.

Completion Checklist

• Project created with metadata
• Team members invited with scoped permissions
• At least one deployment running
• Route configured to active deployment
• Analytics and observability validated
• Guardrails applied

Creation-to-Readiness Flow