Anav Patel contributed to the ggml-org/llama.cpp repository by implementing performance optimizations for Nemotron Nano v2, focusing on enabling CUDA Graph usage to streamline memory copy operations and reduce runtime latency. Leveraging expertise in C++, CUDA, and GPU programming, Anav integrated CUDA Graphs to improve throughput for inference workloads, particularly on edge deployments. The work maintained cross-hardware compatibility while laying the foundation for future graph-based GPU optimizations. Although the contribution spanned a single feature over one month, it demonstrated depth in performance engineering and GPU memory management, directly addressing the need for faster, more efficient Nemotron-based solutions in production environments.