Contributed to the embeddings-benchmark/mteb and huggingface/cookbook repositories by expanding benchmarking capabilities and improving documentation for multilingual users. Developed new benchmark features such as BMRetriever and ReasonIR model integration, enhancing evaluation coverage and model support within the Python-based framework. Addressed metadata accuracy and tokenization reliability, ensuring robust model evaluation and consistent wrapper behavior. Improved onboarding and user experience by refining documentation, correcting Korean language phrasing, and fixing navigation issues in Markdown files. Demonstrated expertise in natural language processing, data curation, and model integration, with a focus on maintainability and accessibility for both users and contributors working in Jupyter Notebook environments.