AI Food Recognition Plugin—Data Collection and Annotation Lead
Led the collection and annotation of over 30,000 real-scene images to train and refine an AI food recognition plugin in the fast-casual restaurant sector. Managed the micro-tuning of visual models for dish occlusion and plate reflection, ensuring enhanced detection accuracy in complex environments. Successfully oversaw data quality and labeling workflows to reach a recognition accuracy of 97.2%. • Built and maintained scalable data pipelines for image annotation tasks. • Collaborated with the algorithm team to set labeling standards and performance metrics. • Implemented iterative feedback processes for label correction and improvement. • Coordinated hardware adaptation based on annotated data outcomes.