Employers have made automated resume screening essential for their recruitment systems because of high competition in the job market today. Previous keyword-based and rule-based methods do not succeed in processing the wide range of resume formats. The proposal research develops an advanced learning system based on YOLOv8 and Tesseract-OCR which detects and extracts unstructured resume documents to create organized data format. The system learns to adapt to any document structure while receiving training with various Personal Details, Education, Experience, Skills layout combinations. The text extraction process begins with Tesseract-OCR being deployed after section detection to transform the text into JSON format. The proposed model reached 91.2% mean average precision at 0.5 precision threshold for section detection and achieved an OCR accuracy of 93.8% using Levenshtein similarity measurements validated on 500+ resume documents. The solution can be effectively implemented at scale because it processes images in 0.94 s per picture on average. Our YOLOv8 + Tesseract system proves to deliver high accuracy at optimal speeds compared to Faster R-CNN, LayoutLMv3 and DocTR. The model demonstrated stability when dealing with both noisy scans and non-standard templates based on the qualitative research and precision-recall plots shown in this work. The research enables scalable automation for recruitment platforms through document understanding systems. Accomplished context awareness in systems could be achieved by implementing multilingual support together with semantic role labeling.