Resume parsing is a technique aimed at extracting relevant information from resumes to enable further processing, such as selection and ranking. Many companies process thousands of resumes during their recruitment processes using traditional methods, such as manual processing and requiring candidates to use a standardized resume template. The current recruitment landscape demands improved approaches in terms of technology and effective methods for resume analysis. Although many basic techniques exist for analyzing structured documents, they are not suitable for unstructured documents (PDF, DOC, DOCX). Current approaches to resume parsing mainly use techniques such as BERT, NLP, keyword-based models, and named entity recognition (NER) models. In this context, this paper proposes an innovative resume parsing system that uses Computer Vision with YOLOv8 and Large Language Models (LLMs), which provide increased accessibility to various APIs. The YOLOv8 model is used for resume segmentation, while Tesseract OCR is employed to extract relevant information in the form of variable text. This information is then processed by two LLMs, integrating the Gemini and OpenAI APIs, which calculate similarity scores and rank candidates based on specific criteria.