I’m building an end-to-end translation management platform on AWS that ingests user-generated content, stores it securely, and runs a series of custom processing workflows before and after Amazon Translate is invoked. The core stack relies on Amazon S3 for object storage, AWS Lambda for all serverless processing, and Amazon Translate for the actual language conversion. Everything needs to be wired together so that new content dropped into S3 automatically triggers the workflow, passes through pre-translation validation and post-translation clean-up, then delivers the final asset back to S3 (or an API endpoint) with relevant metadata and error logs. Key goals • Modular, event-driven architecture with S3 events kicking off Lambda functions • Custom workflow engine inside Lambda layers (or Step Functions if you prefer) where I can slot in additional processors later • Clean IAM boundaries and least-privilege policies • Robust error handling, CloudWatch logging, and retries • Simple status dashboard or API that lets me query job state and retrieve the translated asset Deliverables 1. Infrastructure-as-code (CloudFormation or CDK) for every AWS resource 2. Production-ready Lambda code (Python or Node.js) that handles: – intake validation of user-generated files – streaming requests to Amazon Translate – post-processing (e.g., placeholder preservation, profanity filtering) 3. README with deployment steps, environment variables, and a test script or Postman collection 4. Diagram (draw.io, Lucidchart, or native AWS architecture icon set) showing the full flow Acceptance criteria • Deploys in one command to any new AWS account • Uploading a sample file to the designated S3 bucket returns a translated version plus JSON metadata within 60 seconds • All functions log to CloudWatch with correlation IDs, and failures surface detailed errors while keeping PII masked If you’ve built similar serverless pipelines or have deep experience with S3 events, Lambda concurrency tuning, and Amazon Translate’s quirks, let me know. I’ll share the destination languages and any file-format nuances once we start.