Algorithm Engineer, E-commerce Knowledge Graph - USDS

DescriptionAbout the teamOur team enhances the recognition capabilities for users, products, short videos, and more, in order to recommend more suitable products to users and provide them with an excellent shopping experience.Responsibilities:- Construct user profiling system to accurately depict each user's shopping interests by analyzing their behavioral patterns, preferences, and consumption habits during the shopping process.- Utilize NLP, CV, and multimodal technologies to enhance understanding of short video content and products, supporting search, recommendation, and end-to-end shopping experiences in the marketplace.- Explore large-scale and high-quality pre-training datasets in specific verticals of e-commerce, leveraging the power of the Universal Model to develop industry-specific models for e-commerce.- Discover new scenarios for interactive shopping experiences in e-commerce through these innovations.QualificationsMinimum Qualifications:- Deep research expertise in NLP, CV, and multimodal fields, including but not limited to machine translation, NLP, few-shot learning, reinforcement learning, image and video understanding, and multimodal analysis.- Proficient in training and deploying models using PyTorch, TensorFlow, or other frameworks.- Strong practical skills, with a preference for those who have project experience in video content understanding or multimodal retrieval. Priority will be given to individuals with awards from competitions such as Kaggle, COCO, ActivityNet, ICPC, NOI/IOI.Preferred Qualifications:- Individuals who have published papers in academic conferences/journals- 1+ year relevant work experience from a large-scale internet business.Employment Type: Regular