Design And Development of an AI Powered Open Source GitHub Repository Finder

The GitCompass system offers an effective and intelligent solution for recommending GitHub repositories based on user skills by integrating Natural Language Processing, semantic embedding techniques, and a hybrid ranking strategy. In contrast to traditional methods that depend mainly on keyword matching, the proposed system focuses on understanding the contextual relationship between user skills and repository content, leading to more accurate and meaningful recommendations.   The use of Sentence-BERT for generating semantic embeddings, along with cosine similarity for measuring relevance, enhances the system’s capability to interpret user intent more effectively. Additionally, the hybrid ranking mechanism, which considers multiple factors such as similarity score, repository popularity, and recency of updates, further improves the overall quality and usefulness of the recommendations. The system demonstrates efficient performance with quick response times and consistent behavior across various domains, including machine learning, web development, DevOps, and blockchain.   The evaluation results indicate that the proposed approach performs better than traditional and baseline models in terms of accuracy, relevance, and overall effectiveness. It significantly reduces the manual effort required to search for suitable repositories and simplifies the process of identifying appropriate projects for contribution, especially for beginners.   In conclusion, GitCompass serves as a scalable, reliable, and user-friendly platform that enhances developer productivity and promotes active participation in open-source communities. With future improvements such as enhanced personalization, real-time learning capabilities, and integration with additional platforms, the system has strong potential for practical deployment and continuous advancement.

[1]  J. Kim, J. Wi, and Y. Kim, “Sequential Recommendations on GitHub Repository,” Applied Sciences, vol. 11, no. 4, p. 1585, 2021. DOI: 10.3390/app11041585.   [2]   L. Barhate, P. Chhajed, and S. Vibhute, “AI Powered GitHub Project Management,” International Research Journal of Modernization in Engineering, Technology and Science (IRJMETS), vol. 07, no. 06, pp. 1–6, Jun. 2025. DOI: 10.56726/IRJMETS78747.   [3]    W. Xiao, Y. Sun, D. Lo, and J. Jiang, “RecGFI: Recommending Good First Issues in GitHub OSS Projects,” Proceedings of the IEEE/ACM 44th International Conference on Software Engineering (ICSE), pp. 1–12, 2022.   [4]   G. Mitrov, B. Stanoev, S. Gievska, G. Mirceva, and E. Zdravevski, “Combining Semantic Matching, Word Embeddings, Transformers, and LLMs for Enhanced Document Ranking: Application in Systematic Reviews,” Big Data and Cognitive Computing, vol. 8, no. 9, p. 110, 2024. DOI: 10.3390/bdcc8090110.