Developed and delivered an advanced embedding model integration for the HKUDS/VideoRAG repository, enabling content-based video querying and improved search accuracy. The work focused on incorporating DeepSeek and BAAI/bge-m3 embedding models, establishing end-to-end support for embedding-based retrieval within the VideoRAG project. Leveraging Python, the implementation utilized asynchronous programming and robust API integration to handle complex data flows and video processing tasks. This feature laid the foundation for scalable video search and analytics by supporting advanced embedding models, addressing the need for more precise content discovery and retrieval in large video datasets without introducing any bug fixes during the development period.