TY  - JOUR
AU  - Van, Tran Huu
AU  - Vinh, Nguyen Phu
AU  - Ngan, Vu Thi Tuyet
AU  - Phuong, Luong Ha
AU  - Le, Dac-Nhuong
PY  - 2026
DA  - 2026/05/30
TI  - A Course-Specific Agentic RAG Chatbot for IT Student Support: Architecture, Local Deployment, and Preliminary Evaluation at Hai Phong University
JO  - Next-Generation Computing Systems and Technologies
T2  - Next-Generation Computing Systems and Technologies
JF  - Next-Generation Computing Systems and Technologies
VL  - 2
IS  - 2
SP  - 21
EP  - 34
DO  - 10.62762/NGCST.2026.601800
UR  - https://www.icck.org/article/abs/NGCST.2026.601800
KW  - agentic AI
KW  - educational chatbot
KW  - retrieval-augmented generation
KW  - local large language model
KW  - academic support
KW  - information technology education
AB  - This paper introduces a course-specific agentic retrieval-augmented generation (RAG) chatbot developed to support Information Technology students at Hai Phong University. The proposed system addresses the limitations of static FAQ bots, which are unable to manage open-ended academic tasks, and model-only large language model (LLM) assistants, which may generate fluent yet insufficiently grounded responses. The prototype combines local LLM deployment, semantic retrieval of course materials, and constrained, tool-oriented orchestration to perform four key tasks: question answering, document summarization, study planning, and quiz generation. The primary contributions include a six-layer privacy-aware architecture, a semi-agentic workflow for task routing and source-grounded response generation, and an evaluation protocol that compares LLM-only, RAG, and RAG-plus-orchestration scenarios using a five-dimensional academic rubric. Preliminary local testing demonstrates response times ranging from approximately 2 to 45 seconds for inputs of 1,000-20,000 words, with summary outputs compressed to 10–30% of the original source length. These findings suggest that department-level, course-grounded academic assistance is achievable with modest infrastructure, provided that retrieval, logging, and bounded orchestration are prioritized as core design requirements.
SN  - 3070-3328
PB  - Institute of Central Computation and Knowledge
LA  - English
ER  - 

@article{Van2026A,
  author = {Tran Huu Van and Nguyen Phu Vinh and Vu Thi Tuyet Ngan and Luong Ha Phuong and Dac-Nhuong Le},
  title = {A Course-Specific Agentic RAG Chatbot for IT Student Support: Architecture, Local Deployment, and Preliminary Evaluation at Hai Phong University},
  journal = {Next-Generation Computing Systems and Technologies},
  year = {2026},
  volume = {2},
  number = {2},
  pages = {21-34},
  doi = {10.62762/NGCST.2026.601800},
  url = {https://www.icck.org/article/abs/NGCST.2026.601800},
  abstract = {This paper introduces a course-specific agentic retrieval-augmented generation (RAG) chatbot developed to support Information Technology students at Hai Phong University. The proposed system addresses the limitations of static FAQ bots, which are unable to manage open-ended academic tasks, and model-only large language model (LLM) assistants, which may generate fluent yet insufficiently grounded responses. The prototype combines local LLM deployment, semantic retrieval of course materials, and constrained, tool-oriented orchestration to perform four key tasks: question answering, document summarization, study planning, and quiz generation. The primary contributions include a six-layer privacy-aware architecture, a semi-agentic workflow for task routing and source-grounded response generation, and an evaluation protocol that compares LLM-only, RAG, and RAG-plus-orchestration scenarios using a five-dimensional academic rubric. Preliminary local testing demonstrates response times ranging from approximately 2 to 45 seconds for inputs of 1,000-20,000 words, with summary outputs compressed to 10–30\% of the original source length. These findings suggest that department-level, course-grounded academic assistance is achievable with modest infrastructure, provided that retrieval, logging, and bounded orchestration are prioritized as core design requirements.},
  keywords = {agentic AI, educational chatbot, retrieval-augmented generation, local large language model, academic support, information technology education},
  issn = {3070-3328},
  publisher = {Institute of Central Computation and Knowledge}
}