BACKEND SYSTEM DEVELOPMENT ON MACHINE LEARNING-BASED JOB INTERVIEW APPLICATION AIVUE

• Main Author: Samekto, Rafidika
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/76882
• Institution: Institut Teknologi Bandung

The recruitment process is a process that a candidate must undergo when applying for a job. Often, this process takes a long time, which can waste both the candidate's and the recruiter's time. One of the approaches taken is by using machine learning to expedite a stage in recruitment, and one of these stages is the interview stage. The aspect that can be assisted by machine learning is the emotions displayed by the candidate during the interview. The extraction of these emotions is expected to aid the recruiter's assessment of the candidate because, according to a study, the positive emotions exhibited can bring positive energy to the workplace. Based on these considerations, we have decided to develop AIVue, a machine learning-based job interview application that helps recruiters identify candidate's emotions during the interview. This final project focuses on the development of the application's backend that handles tasks such as communication with the application interface and machine learning models, data storage and retrieval, as well as application development architecture. The result of the development is a backend system of the application that can support the application's operation based on the three previous considerations. Testing of the backend system was conducted by performing integration testing to test each functionality of the defined endpoints. The results of the testing showed that all endpoints can run according to their respective functionalities, but with some findings, namely relatively long response times and occurrences of data inconsistency when using softly deleted data. This final project results in the backend system of the AIVue application that can communicate with the AIVue frontend using APIs, utilizes message queues for machine learning models, and is capable of managing and accessing data stored in a relational database as well as object storage.