RAG ile kurumsal yapay zekâ asistanı geliştirmek: Bir öğrenci destek hizmetleri asistanı örneği

Öz

Kamusal hizmet süreçlerinde veri güvenliği ve kurumsal denetim gereksinimleri, kapalı sistem yapay zekâ çözümlerine olan ilgiyi artırmaktadır. Öğrenci destek hizmetleri gibi kurumsal süreçlerde ticari yapay zekâ modellerinin kullanımı önemli veri güvenliği ve gizlilik riskleri doğurmaktadır. Bu çalışma, "Yapay zekâ araçlarının kamusal alan için özelleştirilmesi ve kapalı bir sistem olarak devreye alınması için hangi süreçler nasıl işletilmelidir?" sorusuna yanıt aramaktadır. Bu amaçla Bilgi Erişimli Üretim (Retrieval-Augmented Generation, RAG) yöntemi ve açık kaynaklı büyük dil modelleri kullanılarak yükseköğretimde bir öğrenci destek hizmetleri yapay zekâ asistanı geliştirilmiştir. Sistem tasarımı, veri hazırlama, model seçimi ile yazılım ve donanım gereksinimleri bütüncül olarak değerlendirilmiştir. Üniversitenin resmî web sitesinden derlenen kurumsal içerik, multilingual-e5-base gömme modeli aracılığıyla Qdrant vektör veri tabanına aktarılmış; Qwen3-4B dil modeliyle bütünleştirilen prototip Gradio arayüzü üzerinden Hugging Face Space platformunda yayınlanmıştır. Çalışma, kamu kurumlarında kuruma özgü yapay zekâ sistemlerinin geliştirilmesine yönelik yol gösterici bir çerçeve sunmaktadır.

Anahtar Kelimeler

• yapay zekâ
• LLM
• açık kaynak
• RAG
• yükseköğretim

Developing an institutional ai assistant with RAG: The case of a student support services assistant

Öz

In public service processes, data security and institutional oversight requirements have increased interest in closed-system artificial intelligence solutions. The use of commercial AI models in institutional processes such as student support services creates significant data security and privacy risks. This study seeks to answer the question: “Which processes should be implemented, and how, in order to customize artificial intelligence tools for the public sector and deploy them as a closed system?” For this purpose, a student support services AI assistant was developed in a higher education context using the Retrieval-Augmented Generation (RAG) method and open-source large language models. System design, data preparation, model selection, and software and hardware requirements were evaluated holistically. Institutional content collected from the university’s official website was transferred to the Qdrant vector database through the multilingual-e5-base embedding model; the prototype, integrated with the Qwen3-4B language model, was published on the Hugging Face Spaces platform through a Gradio interface. The study provides a guiding framework for developing institution-specific AI systems in public institutions.

Anahtar Kelimeler

• artificial intelligence
• LLM
• open source
• RAG
• higher education

Kaynakça

1. Advani, N., Bhat, A. G., Balu-Iyer, S., & Ramanathan, M. (2026). Retrieval augmented generation (RAG) for natural language querying of immunogenicity data for protein drugs. The AAPS Journal, 28:51. doi:10.1208/s12248-025-01199-3
2. Ardic, O., Ozturk, M. U., Demirtas, I., & Arslan, S. (2024). Information extraction from sustainability reports in Turkish through RAG approach. In 2024 32nd Signal Processing and Communications Applications Conference (SIU) (pp. 1–4). IEEE. doi:10.1109/SIU61531.2024.10600994
3. Bahçeci, A. Z., Demirok, F., Kayaoğlu, M., & Topçu, B. (2025). Implementing a RAG-based chatbot in industrial operations: A real-world case study. In 2025 Innovations in Intelligent Systems and Applications Conference (ASYU) (pp. 1–6). IEEE. doi:10.1109/ASYU67174.2025.11208259
4. Brehme, L., Dornauer, B., Ströhle, T., Ehrhart, M., & Breu, R. (2025). Retrieval-augmented generation in industry: An interview study on use cases, requirements, challenges, and evaluation. In Proceedings of the 17th International Joint Conference on Knowledge Discovery, Knowledge Engineering and Knowledge Management - KDIR (pp. 110–122). SciTePress. doi:10.5220/0013739500004000
5. Castaño, J., Martínez-Fernández, S., Franch, X., & Bogner, J. (2024). Analyzing the evolution and maintenance of ML models on Hugging Face. In Proceedings of the 21st International Conference on Mining Software Repositories (pp. 607–618). doi:10.1145/3643991.3644898
6. Çetin, U., & Çolakgil, İ. E. (2025). Türkçe finansal RAG uygulamaları için gömme modellerinin karşılaştırılması. In 2025 33rd Signal Processing and Communications Applications Conference (SIU) (pp. 1–4). IEEE. doi:10.1109/SIU66497.2025.11111893
7. Du, M., Xu, B., Zhu, C., Wang, S., Wang, P., Wang, X., & Mao, Z. (2026). A-RAG: Scaling agentic retrieval-augmented generation via hierarchical retrieval interfaces. arXiv preprint arXiv:2602.03442.
8. Floyd, C. (1984). A systematic look at prototyping. In Approaches to prototyping (pp. 1–18). Berlin, Heidelberg: Springer Berlin Heidelberg.

9. Gao, Y., Xiong, Y., Gao, X., Jia, K., Pan, J., Bi, Y., Dai, Y., Sun, J., Wang, H., & Wang, H. (2023). Retrieval-augmented generation for large language models: A survey. arXiv preprint arXiv:2312.10997.
10. Hartono, L. S., Setiawan, E. I., & Singh, V. (2025). Retrieval augmented generation-based chatbot for prospective and current university students. International Journal of Engineering Science and Information Technology, 5(3), 268–277. doi:10.52088/ijesty.v5i3.951
11. Hevner, A. R., March, S. T., Park, J., & Ram, S. (2004). Design science in information systems research. MIS Quarterly, 28(1), 75–105. doi:10.2307/25148625
12. Husain, M. L., Wibisono, Y., & Anisyah, A. (2025). Development of an academic services chatbot based on retrieval-augmented generation (RAG). Brilliance: Research of Artificial Intelligence, 5(2), 727–735. doi:10.47709/brilliance.v5i2.6719
13. Jain, S. M. (2022). Introduction to transformers for NLP: With the Hugging Face library and models to solve problems. New York, NY: Apress. doi:10.1007/978-1-4842-8844-3
14. Jiang, W., Synovic, N., Hyatt, M., Schorlemmer, T. R., Sethi, R., Lu, Y. H., & Davis, J. C. (2023). An empirical study of pre-trained model reuse in the Hugging Face deep learning model registry. In Proceedings of the 2023 IEEE/ACM 45th International Conference on Software Engineering (ICSE) (pp. 2463–2475). doi:10.1109/ICSE48619.2023.00206
15. Khan, A., Luo, Y., Zhang, W., Zhou, M., & Zhou, X. (2025). Retrieval-augmented generation (RAG): What is there for data management researchers? SIGMOD Record, 54(4), 33–42. doi:10.1145/3793217.3793229
16. Kim, J., & Mahajan, D. (2026). VectorLiteRAG: Latency-aware and fine-grained resource partitioning for efficient RAG. In Proceedings of the 2026 IEEE International Symposium on High Performance Computer Architecture (HPCA) (pp. 1–15). IEEE. doi:10.1109/HPCA68181.2026.11408552
17. Lewis, P., Perez, E., Piktus, A., Petroni, F., Karpukhin, V., Goyal, N., Küttler, H., Lewis, M., Yih, W.-T., Rocktäschel, T., Riedel, S., & Kiela, D. (2020). Retrieval-augmented generation for knowledge-intensive NLP tasks. In Advances in Neural Information Processing Systems 33 (pp. 9459–9474).
18. Liang, C., Cui, Y., Shi, R., Zha, G., Yin, X., Xiao, M., & Huang, B. (2026). GeoAgentic-RAG: A multi-agent framework for autonomous geospatial reasoning and visual insight generation with LLM. International Journal of Applied Earth Observation and Geoinformation, 147, 105195. doi:10.1016/j.jag.2026.105195
19. Liu, B., Zhang, Y., Liu, Y., Deng, X., Qing, J., Han, B., & Qiu, H. (2026). An intelligent multi-modal Q&A system for agriculture combining APGM, PBTCS, and RAG. Smart Agricultural Technology, 13, 101829. doi:10.1016/j.atech.2026.101829
20. Muennighoff, N., Tazi, N., Magne, L., & Reimers, N. (2023). MTEB: Massive text embedding benchmark. In Proceedings of the 17th Conference of the European Chapter of the Association for Computational Linguistics (pp. 2014–2037). Association for Computational Linguistics. doi:10.18653/v1/2023.eacl-main.148
21. Neumann, A. T., Yin, Y., Sowe, S., Decker, S., & Jarke, M. (2024). An LLM-driven chatbot in higher education for databases and information systems. IEEE Transactions on Education, 68(1), 103–116. doi:10.1109/TE.2023.3335520
22. Nunamaker, J. F., Jr., Chen, M., & Purdin, T. D. M. (1990). Systems development in information systems research. Journal of Management Information Systems, 7(3), 89–106. doi:10.1080/07421222.1990.11517898
23. Öner, E. N., Ceyhun, S., Yıldız, M. H., Goncharova, A., Yücel, T. S., Kesgin, H. T., & Amasyali, M. F. (2025). Optimal RAG system design for Turkish textbooks: A comprehensive evaluation and performance enhancement study. In 2025 Innovations in Intelligent Systems and Applications Conference (ASYU) (pp. 1–6). IEEE. doi:10.1109/ASYU67174.2025.11208464
24. Oreški, D., & Vlahek, D. (2024). Retrieval augmented generation in large language models: Development of AI chatbot for student support. In Proceedings for the 15th International Conference on e-Learning 2024 (pp. 45–52). Belgrade, Serbia.
25. Peffers, K., Tuunanen, T., Rothenberger, M. A., & Chatterjee, S. (2007). A design science research methodology for information systems research. Journal of Management Information Systems, 24(3), 45–77. doi:10.2753/MIS0742-1222240302
26. Rodríguez, J. F., Fernández-García, A. J., & Verdú, E. (2026). Grading open-ended questions using LLMs and RAG. Expert Systems, 43(1), e70174. doi:10.1111/exsy.70174
27. Sagynbayeva, A., Pyo, A., Yoon, S. H., & Yang, S. B. (2026). Evaluating user performance on RAG-based generative AI tools: A scenario-based experiment on AI-assisted information retrieval. Computers in Human Behavior. doi:10.1016/j.chb.2026.108952
28. Soliman, H., Kotte, H., Kravčík, M., Pengel, N., & Duong-Trung, N. (2025). Retrieval-augmented chatbots for scalable educational support in higher education. In L. Yan, A. Nguyen, R. Baker, M. Cukurova, D. Gasevic, K. Yang, ... L. Zhao (Eds.), CEUR Workshop Proceedings: Vol. 3994. Proceedings of the Second International Workshop on Generative AI for Learning Analytics (GenAI-LA 2025) (pp. 22–31). Aachen, Germany: CEUR-WS.org.
29. Taneja, K., Maiti, P., Kakar, S., Guruprasad, P., Rao, S., & Goel, A. K. (2024). Jill watson: A virtual teaching assistant powered by chatgpt. In International Conference on Artificial Intelligence in Education (pp. 324–337). Cham, Switzerland: Springer Nature Switzerland.
30. Tekgöz, H., Erdoğan, H., Uz, H., & Altınok, H. (2025). Evaluation of vector database and LLM models in Retrieval-Augmented Generation (RAG) systems. In 2025 10th International Conference on Computer Science and Engineering (UBMK) (pp. 572–576). IEEE. doi: 10.1109/UBMK67458.2025.11206778
31. Wang, F., & Hannafin, M. J. (2005). Design-based research and technology-enhanced learning environments. Educational Technology Research and Development, 53(4), 5–23. doi:10.1007/BF02504682
32. Wu, J.-W., & Tseng, M.-H. (2025). Developing a local generative AI teaching assistant system: Utilizing Retrieval-Augmented Generation technology to enhance the campus learning environment. Electronics, 14(17), 3402. doi:10.3390/electronics14173402
33. Xu, J., Xu, Z., Jiang, Z., Chen, Z., Luo, H., Wang, Y., & Gui, W. (2026). Labeling-free RAG-enhanced LLM for intelligent fault diagnosis via reinforcement learning. Advanced Engineering Informatics, 69, 103864. doi:10.1016/j.aei.2025.103864