Worked on enhancing backup reliability in the wal-g/wal-g repository by improving S3 multipart upload processes. Addressed issues where transient HTTP 409 errors and network timeouts caused backup failures, particularly during concurrent operations or on unstable networks. Implemented robust retry mechanisms and refined error handling in Go, leveraging AWS S3 integration and backend development expertise. Adjusted logging to reduce unnecessary warnings while maintaining operator visibility, supporting better observability and operational clarity. Collaborated on solutions aligned with reported issues, demonstrating attention to concurrency and error resilience. The work emphasized network programming, unit testing, and practical improvements to backup availability in distributed environments.