Best way to batch process rows from a table with LLM - using AWS Bedrock

Beginners
1

Does anyone know what is the best way to batch process a bunch of rows with an LLM?

I am using AWS bedrock. Basically i have a DynamoDB table and I have an LLM prompt to run for each row. I need to return a JSON object with status: approved | rejected and an note - this is not an answer evaluation job but more of a classification.

**
What is the best way to process set this pipeline - considering cost and resources. I know AWS has a batch inference endpoint.**

1 Like
2

Seems like it’ll don’t get dramatically cheaper, though…