Authors:
Ojas M. Agarwal, Devi Karuppiah, Lekshmi Kalinathan, Benila Sathianesan, Monish Sudhagar, Vikram Ramkumar, Advaith T. Ajith

Addresses:
School of Computer Science and Engineering, Vellore Institute of Technology, Chennai, Tamil Nadu, India.

Abstract:

This paper reviews a novel LLM-powered pipeline for generating resumes that are dynamically tailored to job descriptions and optimised for Applicant Tracking Systems (ATS). In today’s hiring landscape, most resumes are filtered before reaching a recruiter, primarily because of the rigid keyword-matching and formatting rules enforced by ATS algorithms. This presents a significant challenge for job seekers, especially those who cannot tailor their resumes for each application. Traditional resume tools often offer static templates or simple keyword suggestions, but generally fail to contextualise content and provide useful feedback. By combining structured parsing, semantic alignment, and iterative scoring, our system overcomes these limitations and generates highly customised, ATS-friendly resumes with minimal user effort. We demonstrate, through token-based and latent-space evaluations, that our optimised resumes nearly double their effectiveness compared to unmodified versions, improving alignment scores on average by 85%. This presents our approach as a valuable tool that enables job seekers to compete in increasingly competitive markets, while also representing a technological advancement.

Keywords: Large Language Model (LLM); Applicant Tracking System (ATS); AI Resume Generation; Semantic Alignment; Job Description Matching; ATS Algorithms; Technological Advancement.

Received on: 23/11/2024, Revised on: 17/02/2025, Accepted on: 28/03/2025, Published on: 07/09/2025

DOI: 10.69888/FTSCS.2025.000477

FMDB Transactions on Sustainable Computing Systems, 2025 Vol. 3 No. 3, Pages: 192-202

• Views : 74
• Downloads : 7

Download PDF