Worked on enhancing data reliability and multilingual benchmarking in open-source research repositories. In acl-org/acl-anthology, addressed metadata integrity by correcting conference file paths and updating references, which improved the accuracy of downstream indexing and citation pipelines for researchers. Later, contributed to embeddings-benchmark/mteb by integrating a Multilingual News Article Retrieval Dataset, expanding the repository’s language coverage and updating dataset statistics and metadata for robust evaluation. Employed Python for data analysis, metadata management, and machine learning tasks, focusing on precise documentation and validation. The work emphasized careful alignment of metadata and datasets to ensure reliable search, indexing, and benchmarking across projects.