chore: Cleanup AWS SageMaker AI docs, examples, and provider

[will-holley]

Changes

• Migrates away from using hardcoded model-specific request payload serialization logic to requiring that the user provide a request serializer which handles both the input and model parameters. This requires the transform property and drops support for individual model settings on the config parameter. This change unlocks the ability to use any SageMaker Jumpstart or custom model.

Fixes

• Renames "SageMaker" to "SageMaker AI" within the documentation and example. SageMaker AI is distinct from SageMaker and the service which our provider interfaces with.
• Removes Anthropic from provider; it's on Bedrock, not SageMaker.

[use-tusk]

⏩ No test execution environment matched (5951e01) View output ↗

View output in GitHub ↗

Tip

New to Tusk? Learn more here.

View check history

Commit	Status	Output	Created (UTC)
1459893	⏩ No test execution environment matched	Output	Jun 27, 2025 3:03PM
f70a572	⏩ No test execution environment matched	Output	Jun 27, 2025 5:00PM
5951e01	⏩ No test execution environment matched	Output	Jun 27, 2025 5:02PM