AI Intern – Car Damage Detection Data Labeling & Model Training
Researched and trained YOLO computer vision models for an AI-driven car damage detection system for insurance companies. Optimized object detection models using over 8,000 labeled images, focusing on various car damage categories. Utilized AWS EC2 and S3 for large-scale data deployment and cloud-based processing. • Annotated image datasets for car damage using bounding boxes or segmentation. • Validated the accuracy of AI-generated labels by comparing them with ground-truth annotations. • Curated and selected representative images to improve model generalizability. • Collaborated with a technical team to align labeling outputs with project goals.