What problem does it solve? Large data exports block API requests and degrade the user experience. This Skill implements an asynchronous export pipeline where POST /api/export returns a job_id immediately while heavy Polars joins and serialization run in BackgroundTasks, keeping the rest of the API responsive. ## Core Features & Use Cases - Non-blocking export endpoint: POST /api/export validates the request, registers the job, and returns 202 with a job_id instantly; the worker runs in FastAPI BackgroundTasks. - Job state tracking: Jobs move through queued, running, done, and failed states persisted in the export_job table, with audit fields for user, filters, size, and duration. - GCS delivery with signed URLs: Finished CSV/XLSX files upload to GCS under exports/{user}/{job_id} with 24-hour signed URLs and a 7-day bucket lifecycle. - Use Case: An analyst requests a 1-million-row export as XLSX; the API responds immediately, the worker builds the file via the semantic layer, and the analyst polls GET /api/export/{job_id} until a signed download URL appears. ## Quick Start Ask the AI to implement the export endpoint and background worker so that POST /api/export returns a job_id immediately and the finished file is uploaded to GCS with a signed URL.
