Worked on the ofiwg/libfabric repository to optimize concurrency and memory management in C, focusing on improving performance and scalability for high-concurrency workloads. Refactored memory operations by dropping the memory-monitor lock before executing memory tasks and pre-allocating red-black tree nodes outside the lock, which reduced lock contention and mitigated potential deadlocks. Enhanced memory region entry handling to prevent freeing valid entries when the cache is full or a subscription fails. Made allocation and deallocation functions public and simplified internal logic based on maintainer feedback. These changes improved throughput, reduced latency, and increased stability under concurrent workloads using C and concurrent programming.