ARCHIVES

The rapid growth of digital recruitment platforms has resulted in a massive volume of resumes for each job opening, making manual candidate screening time-consuming, inconsistent and difficult to scale. Automated hiring tools are increasingly required to help organizations shortlist suitable candidates quickly while maintaining fairness, transparency and alignment with job requirements. This paper presents an autonomous AI hiring agent that combines natural language processing, Large Language Models (LLMs), classical machine learning and graph-based modeling to support data-driven hiring decisions. The proposed system automatically parses resumes, extracts relevant technical skills and experience, and represents both candidates and job descriptions as structured feature vectors. A TF-IDF and cosine-similarity based matching engine is used to compute relevance scores between each candidate and multiple job roles, which are then represented in a candidate–job graph where nodes correspond to entities and weighted edges indicate match strength. On top of this graph, decision rules and configurable thresholds categorize candidates into levels such as strong hire, consider or reject, thereby imitating the behavior of an intelligent recruiter or modern Applicant Tracking System. The system is implemented using Python, scikit-learn, NetworkX and a Streamlit-based interface, enabling interactive visualization of matches and easy interpretation for HR users. The architecture is modular and can be extended with domain-specific LLMs, knowledge-graph integrations and real-time labour-market data, making it suitable for academic study as well as practical use in AI-driven recruitment automation.