Worked on the ydb-platform/ydb-go-sdk repository to address a critical race condition in the stream reconnector, focusing on improving the reliability of long-running streaming operations. Applied Go concurrency patterns and careful mutex management to ensure that stream operations occurred outside the mutex lock, preventing deadlocks and data corruption. Refactored the CloseWithError method to safely manage the lifecycle of streaming resources, enhancing error handling and concurrency safety. This work stabilized the streaming path under reconnection scenarios, aligning with project resilience goals. Demonstrated strong skills in Go, concurrency, and refactoring, delivering measurable improvements to code quality and operational robustness within the codebase.