Developed and integrated ZBAL support for NPU memory management within the yhyang201/sglang repository, focusing on distributed processing environments. The work centered on enabling optimized memory allocation for NPUs by incorporating the Zero Buffer Accelerate Library, with configuration managed through environment variables and dedicated initialization hooks. This approach allowed seamless integration of ZBAL into existing NPU workflows, improving memory efficiency and scalability for distributed workloads. The implementation was delivered in Python, demonstrating skills in distributed systems, memory management, and NPU development. No major bugs were addressed during this period, with efforts concentrated on foundational feature development for future scalability.