A data scientist stores financial datasets in S3 and queries them with Athena. They deployed an ML model on a SageMaker endpoint and want to invoke it from Athena, but invocations fail and the data scientist's IAM user cannot call the endpoint. Which combination of actions will enable the IAM user to invoke the SageMaker endpoint? (Choose three.)

To invoke a SageMaker endpoint from Athena, the IAM user running the Athena query needs explicit permission to call the SageMaker endpoint. This is granted by adding an IAM policy statement allowing the sagemaker:InvokeEndpoint action. Athena also needs to be configured to use the external ML function, which is done by including the USING EXTERNAL FUNCTION mlfunctionname clause in the SQL query. Finally, if the SageMaker model needs to access data in S3 (e.g., for feature engineering or inference data), the SageMaker execution role must have permissions to read from the relevant S3 buckets. "Attach the AmazonAthenaFullAccess managed policy to the IAM user" is incorrect because while it grants full Athena access, it does not include permissions to invoke SageMaker endpoints. "Add an IAM policy statement that allows the user to perform the sagemaker:GetRecord action" is incorrect as GetRecord is for Feature Store, not endpoint invocation. "Remap the IAM user in SageMaker to a different IAM user that has access to the hosted endpoint" is not a valid or necessary action; permissions are managed via IAM policies.