Worked on the m1a2st/kafka repository to address reliability issues in Kafka log management, focusing on concurrency safety during log deletion. Investigated and resolved a race condition that could prematurely close the source log, potentially taking log directories offline during concurrent flush and delete operations. The solution involved careful concurrency debugging and Scala backend development, ensuring that log lifecycle events are handled safely under parallel workloads. Collaborated with reviewers to validate the fix, documented the approach with Jira references, and contributed to release engineering. This work improved data integrity and production uptime by safeguarding log directories against outages during concurrent operations.