Developed an OpenVINO Optimization Toolkit for the open-edge-platform/edge-ai-suites repository, focusing on enhancing the Pi0.5 Vision-Language-Action model for efficient on-device inference. The work involved creating and documenting scripts in Python to automate model conversion, weight compression, and benchmarking, enabling reproducible optimization workflows. Leveraging expertise in machine learning, model optimization, and robotics, the developer established a clear process for deploying optimized models on Pi0.5 hardware. Collaboration with Intel/OpenVINO was strengthened to support future hardware-optimized deployments. The contribution provided comprehensive documentation and practical guidance, streamlining the integration of OpenVINO-optimized models into edge AI applications for robotics platforms.