Methods of semantic analysis of competencies in intelligent recruiting systems based on large language models

Abstract

The article investigates methods for transforming unstructured text data of resumes and job descriptions into semantic vectors of competencies for intelligent recruiting tasks. The evolution of approaches from classical statistical methods of natural language processing to modern transformer architectures based on large language models (LLM) is considered. A comparative analysis of existing recruiting platforms is conducted and the concept of the author's system that uses the OpenAI API for Named Entity Recognition (NER) with transparent scoring of the candidate's suitability for the job is substantiated. Special attention is paid to the problems of morphosyntactic analysis of the Ukrainian language and methods of its lemmatization