Worked on strengthening diagnostics and error handling in distributed systems, focusing on the EFA driver stack within the amzn/amzn-drivers and ofiwg/libfabric repositories. Addressed two targeted bugs by implementing explicit error handling for feature mismatches during peer communication and RDMA operations, using C programming and kernel development skills. Enhanced diagnostic clarity by integrating consistent error messaging and errno handling, which improved maintainability and reduced troubleshooting time for networked workloads. The approach emphasized robust feedback when peers lacked requested features, supporting more reliable feature negotiation across modules. This work contributed to improved system reliability and facilitated smoother cross-repository collaboration for distributed environments.