Contributed backend features and performance optimizations across vllm-project/production-stack, LMCache, vllm, and MystenLabs/sui, focusing on asynchronous programming, concurrency, and caching. Enhanced routing efficiency in production-stack by implementing round-robin endpoint caching and improved concurrent data structures with asynchronous locks. In LMCache, accelerated cache lookups by replacing SHA256 with xxhash, optimizing dependency management in Python. Refactored cache manager initialization and streamlined unit testing setup in vllm to support contributor onboarding. For MystenLabs/sui, developed an Admin Tracing Server in Rust, introducing optional tracing for robust error handling and improved observability, supporting advanced workflows like force-close-epoch in distributed systems.