Laporan kerja praktek I 18 Februari–12 Juni 2026 di PT. Sentra Vidya Utama (SEVIMA)

Prasetya, Rahman Azkarafi (2026) Laporan kerja praktek I 18 Februari–12 Juni 2026 di PT. Sentra Vidya Utama (SEVIMA). Project Report. [s.n], [s.l.]. (Unpublished)

Text 5025231044-Project_Report.pdf - Accepted Version Restricted to Repository staff only Download (2MB) | Request a copy

Abstract

Platform pembelajaran digital saat ini umumnya menyimpan materi perkuliahan dalam format dokumen pasif seperti PDF dan PowerPoint. Mahasiswa seringkali menghadapi kendala dalam menelaah materi yang ekstensif secara mandiri, sementara dosen memiliki keterbatasan waktu untuk memberikan bimbingan personal secara real-time. Kesenjangan ini mendorong perlunya transformasi repositori dokumen menjadi basis pengetahuan aktif yang mampu merespons kebutuhan mahasiswa secara cerdas dan kontekstual. Kerja praktik ini bertujuan untuk mengembangkan Sevima RAG Hub, sebuah infrastruktur kecerdasan buatan berbasis Retrieval-Augmented Generation (RAG) yang dirancang sebagai layanan terpusat untuk pemrosesan informasi akademik. Sistem dibangun menggunakan arsitektur backend microservices berbasis FastAPI yang terintegrasi dengan Vector Database untuk penyimpanan embedding dan Cloud LLM API untuk inferensi jawaban. Pengembangan dilaksanakan dalam enam siklus sprint yang mencakup penelitian pipeline AI, implementasi backend, pengembangan frontend, pengujian fungsional, serta evaluasi kualitas jawaban AI. Hasil pengujian performa menunjukkan skor keseluruhan rata-rata mencapai 90,33% dengan skor faithfulness 100%, yang berarti seluruh jawaban yang dihasilkan sepenuhnya bersumber dari dokumen tanpa halusinasi. Sistem berhasil menyediakan API terstandarisasi yang terdokumentasi melalui Swagger UI, antarmuka fungsional untuk dosen dan mahasiswa, serta widget embed yang memungkinkan integrasi dengan sistem eksternal melalui iframe.
==================================================================================================================================
Current digital learning platforms generally store course materials in passive document formats such as PDF and PowerPoint. Students often encounter difficulties in studying extensive learning materials independently, while lecturers have limited time to provide personalized guidance in real time. This gap highlights the need to transform document repositories into active knowledge bases capable of delivering intelligent and context-aware responses to students' academic needs. This internship aims to develop Sevima RAG Hub, an artificial intelligence infrastructure based on Retrieval-Augmented Generation (RAG) designed as a centralized service for academic information processing. The system is built using a FastAPI-based microservices backend architecture integrated with a vector database for embedding storage and a cloud-based Large Language Model (LLM) API for answer generation. The development process was carried out through six sprint cycles, covering AI pipeline research, backend implementation, frontend development, functional testing, and evaluation of AI-generated response quality. The performance evaluation results show an overall average score of 90.33%, with a faithfulness score of 100%, indicating that all generated responses were fully grounded in the source documents without hallucinations. The system successfully provides standardized APIs documented through Swagger UI, functional interfaces for lecturers and students, and an embeddable widget that enables integration with external systems through an iframe.

Item Type:	Monograph (Project Report)
Uncontrolled Keywords:	Retrieval-Augmented Generation, Microservices, FastAPI, Knowledge Base, Intelligent Learning Assistant
Subjects:	Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines. Q Science > QA Mathematics > QA76.754 Software architecture. Computer software
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Informatics Engineering > 55201-(S1) Undergraduate Thesis
Depositing User:	Rahman Azkarafi Prasetya
Date Deposited:	29 Jun 2026 02:09
Last Modified:	29 Jun 2026 02:09
URI:	http://repository.its.ac.id/id/eprint/134118

Actions (login required)

View Item