Phase 6: Implement retries with exponential backoff

Ensure failed jobs retry gracefully with backoff strategy.

The issue mentions implementing retries with exponential backoff for failed jobs, likely in the worker layer using BullMQ. Start by examining the existing job queue setup in the workers directory, understand how jobs are currently processed and where failures occur. Look for configuration related to retry logic in BullMQ or the job processor. 'Done' means jobs automatically retry with increasing delays after failures.

Written by the indexing model from the issue text.