Developed a batch-capable blocking retrieval path for the LMCache repository’s NixlStorageBackend, focusing on synchronous access to cached memory objects. The work involved refactoring existing backend methods to optimize locking mechanisms and memory allocation, resulting in more predictable memory usage and safer concurrent access patterns. Leveraging Python and applying skills in asynchronous programming, backend development, and memory management, the developer enabled reliable synchronous workloads and improved maintainability of the storage backend. These enhancements support better performance characterization under load and establish a foundation for future backend improvements, directly aligning with the platform’s caching strategy and requirements for robust memory-object retrieval.