PEMANFAATAN KECERDASAN BUATAN DAN SPEECH RECOGNITION UNTUK APLIKASI LATIHAN WAWANCARA KERJA SOFTWARE ENGINEER BERBASIS WEB

Farid Rizky Wijaya; Nunik Pratiwi;

ABSTRAK

Pemanfaatan Kecerdasan Buatan dan Speech Recognition untuk Aplikasi Latihan Wawancara Kerja Software Engineer Berbasis Web

Farid Rizky Wijaya

Persaingan kerja di bidang software engineering semakin ketat seiring meningkatnya jumlah
programmer dan tingginya tuntutan industri terhadap kesiapan kandidat, tidak hanya dari sisi
kemampuan teknis tetapi juga soft skills seperti komunikasi, khususnya dalam proses wawancara
kerja. Namun, banyak calon software engineer masih mengalami kesulitan dalam menyampaikan
pengalaman, menjawab pertanyaan secara terstruktur, serta merasa kurang percaya diri saat
menghadapi wawancara kerja. Oleh karena itu, diperlukan media latihan wawancara yang interaktif
dan mudah diakses untuk membantu meningkatkan kesiapan calon software engineer. Penelitian ini
bertujuan untuk membuat aplikasi latihan wawancara berbasis web dengan memanfaatkan
kecerdasan buatan (Artificial Intelligence) dan teknologi pengenalan ucapan (speech recognition).
Aplikasi ini mengintegrasikan Large Language Model (LLM) Gemini AI dengan teknik wawancara
Behavioral Event Interview (BEI) serta teknik prompt chaining untuk menghasilkan simulasi
wawancara yang kontekstual dan berkelanjutan. Metode pengembangan sistem yang digunakan
adalah Personal Extreme Programming (PXP) yang bersifat iteratif, meliputi tahap perencanaan,
desain, implementasi, pengujian, dan refleksi. Pengujian fungsional dilakukan menggunakan Black
Box Testing, sedangkan evaluasi usability dilakukan menggunakan System Usability Scale (SUS).
Hasil evaluasi menunjukkan nilai rata-rata SUS sebesar 82,36, yang termasuk dalam Grade A,
sehingga mengindikasikan bahwa aplikasi memiliki tingkat kemudahan penggunaan dan
penerimaan pengguna yang sangat baik.

Kata kunci: Large Language Model, Pengenalan Ucapan, Latihan Wawancara Kerja, Personal Extreme Programming.

ABSTRACT

Utilization of Artificial Intelligence and Speech Recognition for a Web-Based Job Interview Training Application for Software Engineers

Farid Rizky Wijaya

Competition in the software engineering job market has become increasingly intense due to the
growing number of programmers and the high demands of industry for candidate readiness, not only
in terms of technical skills but also soft skills such as communication, particularly in the job
interview process. However, many prospective software engineers still experience difficulties in
conveying their experiences, answering interview questions in a structured manner, and maintaining
confidence during interviews. Therefore, an interactive and easily accessible interview practice
medium is needed to help improve the readiness of prospective software engineers. This study aims
to build a web-based interview practice application by utilizing Artificial Intelligence (AI) and
speech recognition technology. The application integrates the Gemini AI Large Language Model
(LLM) with the Behavioral Event Interview (BEI) technique and prompt chaining to generate
contextual and continuous interview simulations. The system was developed using the Personal
Extreme Programming (PXP) methodology, which follows an iterative process consisting of
planning, design, implementation, testing, and reflection. Functional testing was conducted using
Black Box Testing, while usability evaluation was carried out using the System Usability Scale (SUS).
The evaluation results show an average SUS score of 82.36, which falls into Grade A, indicating
that the application has a high level of usability and strong user acceptance.

Keyword: Large Language Model, Speech Recognition, Job Interview Training, Personal Extreme Programming.