Research on Intelligent Competency Modeling and Job-Person Matching Mechanism Based on the DeepSeek Model

With the rapid development of artificial intelligence, especially open-source large language models, enterprise human resource management is undergoing profound transformation. Traditional competency models for job positions largely rely on expert experience for construction, which leads to outdated updates, poor generalizability, and an inability to meet the rapidly changing strategic needs of modern organizations [1]. This paper proposes a low-cost, scalable competency modeling and job-person matching framework based on DeepSeek, a representative open-source language model. By applying natural language processing and semantic embedding techniques, the system extracts competency elements from job descriptions and resumes to construct ability vectors and compute job-person matching scores. The study also explores the application of this model in scenarios such as recruitment screening, job adjustment, and talent development, and identifies challenges such as data quality, model interpretability, and digital literacy among HR professionals, while proposing corresponding countermeasures. Experiments and preliminary applications demonstrate that the intelligent competency system based on large language models is highly feasible and commercially valuable [2].