You can add the ARN of an Inference profile to your modelID itself while invoking the model.

response = bedrock_client.invoke_model(

       modelId="arn:aws:bedrock:us-east-1::model/your-bedrock-model-arn"  

       prompt="Your prompt here"

   )