During July 2025, Tingqian Li contributed to the openvinotoolkit/openvino.genai repository by optimizing the VisionEncoderQwen2VL image encoding pipeline. Li refactored the preprocessing steps—resize, normalization, and reshape—directly into the OpenVINO model graph using C++, enabling hardware acceleration and reducing unnecessary data movement. This approach improved image encoding throughput and lowered latency, enhancing end-to-end inference performance for downstream computer vision workflows. The work demonstrated strong skills in C++ development, model optimization, and integration with OpenVINO. While the contribution focused on a single feature, it addressed a core performance bottleneck and reflected a deep understanding of model-graph engineering.