Pembuatan Portal Kerja Cerdas dengan Fitur Pencocokan Kandidat dan Applicant Tracking System

Abstract

Proses rekrutmen konvensional, khususnya bagi Usaha Mikro, Kecil, dan Menengah (UMKM), masih menghadapi kendala seperti volume lamaran tidak relevan, penyaringan manual yang tidak efisien dan rentan bias, serta minimnya transparansi status lamaran bagi pencari kerja. Penelitian ini bertujuan merancang dan membangun Portal Kerja Cerdas "Talentara" berbasis web yang mengintegrasikan Applicant Tracking System (ATS), sistem pencocokan kandidat berbasis rule-based system, halaman karier perusahaan, notifikasi status lamaran secara real-time, dan fasilitasi wawancara video. Sistem dikembangkan secara iteratif menggunakan metodologi Agile Kanban selama delapan bulan dengan batas Work In Progress (WIP) tiga kartu, dibangun menggunakan Laravel berarsitektur Model-View-Controller. Algoritma pencocokan kandidat menggunakan pendekatan multifaktor berbobot yang mempertimbangkan kecocokan keahlian, pengalaman kerja, riwayat pendidikan, lokasi, dan posisi, sedangkan pengujian fungsional dilakukan dengan Black Box Testing menggunakan Equivalence Partitioning dan Boundary Value Analysis. Sistem berhasil mengakomodasi empat peran pengguna (Kandidat, Perusahaan, Admin, Pemilik) dengan dasbor dan fungsionalitas masing-masing. Hasil pengujian terhadap 193 kasus uji menunjukkan tingkat keberhasilan 96,4%, dengan tujuh kasus gagal yang telah diperbaiki. Sistem pencocokan kandidat memberikan rekomendasi dua arah, sementara ATS memungkinkan pengelolaan alur seleksi kandidat secara terstruktur, sehingga Talentara dinilai mampu mendukung proses rekrutmen digital yang lebih terstruktur, efisien, dan transparan bagi UMKM.

Conventional recruitment processes, particularly for Micro, Small, and Medium Enterprises (MSMEs), still face challenges such as a high volume of irrelevant applications, inefficient and bias-prone manual screening, and limited transparency in application status for job seekers. This study aims to design and build "Talentara," a web-based Smart Job Portal integrating an Applicant Tracking System (ATS), a rule-based candidate matching system, company career pages, real-time application status notifications, and video interview facilitation. The system was developed iteratively using the Agile Kanban methodology over eight months with a Work In Progress (WIP) limit of three cards, built using Laravel with a Model-View-Controller architecture. The matching algorithm employs a weighted multi-factor approach considering skill, experience, education, location, and position relevance, while functional testing used Black Box Testing with Equivalence Partitioning and Boundary Value Analysis. The system accommodates four user roles (Candidate, Company, Admin, Owner), each with its own dashboard and functionalities. Testing on 193 test cases showed a 96.4% success rate, with seven failed cases subsequently fixed. The matching system provides bidirectional recommendations, while the ATS enables a structured candidate selection workflow, making Talentara capable of supporting a more structured, efficient, and transparent digital recruitment process for MSMEs.