Athaillah, Alendra Rafif (2025) Implementasi Sistem Inferensi Real-Time Berbasis Arsitektur Serverless Menggunakan AWS Lambda dan DynamoDB pada Data Customer di PT. Telekomunikasi Seluler. Project Report. [s.n], [s.l.]. (Unpublished)
![[thumbnail of 5025221297-Project_Report.pdf]](http://repository.its.ac.id/style/images/fileicons/text.png) |
Text
5025221297-Project_Report.pdf - Accepted Version Restricted to Repository staff only Download (936kB) | Request a copy |
Abstract
Dalam industri telekomunikasi digital yang kompetitif, kemampuan perusahaan dalam memanfaatkan data pelanggan untuk mendukung pengambilan keputusan menjadi faktor penting dalam meningkatkan kualitas layanan dan efisiensi bisnis. Salah satu pendekatan yang dapat digunakan adalah penerapan sistem inferensi real-time untuk memprediksi pengeluaran bulanan pelanggan sebagai dasar dalam pemberian rekomendasi produk dan strategi pemasaran. Proyek kerja praktik ini bertujuan untuk mengembangkan sistem inferensi real-time berbasis serverless architecture menggunakan layanan Amazon Web Services (AWS). Sistem dibangun dengan komponen utama AWS Lambda sebagai lingkungan eksekusi model, Amazon DynamoDB sebagai penyimpanan data pelanggan, Amazon API Gateway sebagai antarmuka REST API, serta Amazon S3 untuk penyimpanan model dan dependensi. Model yang digunakan dilatih menggunakan algoritma CatBoost Regressor. Hasil implementasi menunjukkan bahwa sistem mampu memberikan hasil prediksi secara real-time dengan rata-rata latensi sebesar 106,6 ms, latensi terendah 11,19 ms, dan tertinggi 260,31 ms. Performa tersebut menunjukkan bahwa arsitektur serverless efektif dalam mendukung kebutuhan inferensi cepat dan efisien tanpa memerlukan pengelolaan infrastruktur server secara langsung. Sistem ini berpotensi untuk diintegrasikan lebih lanjut dengan sistem rekomendasi produk atau analitik pelanggan dalam skala produksi.
====================================================================================================================================
In the highly competitive digital telecommunications industry, a company’s ability to leverage customer data to support decision-making is a crucial factor in improving service quality and business efficiency. One approach that can be applied is the implementation of a real-time inference system to predict customers’ monthly expenditures as a basis for product recommendations and marketing strategies.This internship project aims to develop a real-time inference system based on a serverless architecture using Amazon Web Services (AWS). The system is built with AWS Lambda as the model execution environment, Amazon DynamoDB for customer data storage, Amazon API Gateway as the REST API interface, and Amazon S3 for storing the trained model and its dependencies. The predictive model is trained using the CatBoost Regressor algorithm.The implementation results show that the system is capable of delivering real-time predictions with an average latency of 106.6 ms, a minimum latency of 11.19 ms, and a maximum latency of 260.31 ms. This performance demonstrates that a serverless architecture is effective in supporting fast and efficient inference without the need for direct server infrastructure management. The system has strong potential to be further integrated with product recommendation systems or customer analytics in a production-scale environment.
| Item Type: | Monograph (Project Report) |
|---|---|
| Uncontrolled Keywords: | Serverless Architecture, AWS Lambda, DynamoDB, API Gateway, CatBoost, Real-Time Inference |
| Subjects: | Q Science > QA Mathematics > QA336 Artificial Intelligence Q Science > QA Mathematics > QA76 Computer software Q Science > QA Mathematics > QA76.585 Cloud computing. Mobile computing. Q Science > QA Mathematics > QA76.9.C55 Client/server computing |
| Divisions: | Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Informatics Engineering > 55201-(S1) Undergraduate Thesis |
| Depositing User: | Alendra Rafif Athaillah |
| Date Deposited: | 17 Dec 2025 06:43 |
| Last Modified: | 17 Dec 2025 06:43 |
| URI: | http://repository.its.ac.id/id/eprint/129029 |
Actions (login required)
 |
View Item |