How do you design multi-agent orchestration for document processing?hard
Answer
Orchestrator agent delegates to: parsing agent (extract raw text from PDF), classification agent (identify document type/section), metadata enrichment agent (extract structured financial fields), and validation agent (sc
Explanation
Orchestrator agent delegates to: parsing agent (extract raw text from PDF), classification agent (identify document type/section), metadata enrichment agent (extract structured financial fields), and validation agent (schema compliance). Step Functions orchestrates the state machine — each agent is a Lambda function. EventBridge triggers pipeline on S3 uploads. Results written to DynamoDB.
Follow-upWhat tradeoffs did you consider in that implementation?