Worked on the htcondor/htcondor repository to enhance scheduler reliability in high-concurrency environments by addressing a critical bug in concurrency limits evaluation and claim recycling. Focused on ensuring that partitionable slot leftovers were correctly managed, the developer improved resource accounting and prevented stale claims from impacting scheduling decisions. Using C++ and leveraging expertise in concurrency control and system programming, the solution stabilized scheduling behavior in edge cases and contributed to more predictable throughput for busy clusters. The work was delivered as a focused bug fix, with clear traceability through a dedicated commit and changelog entry, reflecting a methodical debugging approach.