Worked on enhancing the stability of large corpus downloads in the google/clusterfuzz repository by addressing memory and corpus management challenges. Focused on increasing the BOT_TMPDIR memory allocation to 1GB and refactoring the corpus synchronization process to utilize a dedicated fuzzer-testcases directory with a 1.25GB tmpfs mount. These changes, implemented using Python and shell scripting within a Linux environment, reduced the occurrence of “No space left on device” errors during large-scale corpus synchronization. The improvements enabled more reliable handling of large corpora without requiring infrastructure changes, demonstrating skills in DevOps, containerization, and memory management for backend systems.