The main business scenarios are user growth, interactive shopping guides, message recommendations, etc. of Taobao Mobile, including external advertising algorithms for products and content, personalized recommendations, Push message recommendations, special editions and other algorithm scenarios. 1. Upgrade user portraits and create a Taobao mobile user growth delivery engine, including product material recall, sorting, and LT/LTV modeling 2. Build personalized recommendation models for thousands of scenarios, and be responsible for the design and efficiency improvement of large model migration rely on the personalized recommendation platform to support recommendation algorithms for thousands of recommendation scenarios and sorting models with tens of billions of parameters. Deeply understand the business needs of recommendation, design and implement large-scale recall/sorting/rearrangement algorithms, and continuously improve the recommendation effect and user experience 3. Design uplift models for red envelopes and coupons and other rights demand scenarios to improve incremental indicators 4. Deeply understand message recommendation, continuously optimize user experience and improve core indicators through timing algorithms, recall, sorting, closing rate models, etc. 5. Improve algorithm efficiency through sequence modeling, transfer learning, graph model, multi-crowd multi-target modeling and other technologies,1. Computer science or related majors have solid data structure, algorithm and coding skills, and be proficient in at least one programming language, such as C/C++, JAVA, Python, etc. 2. Solid and familiar with the basics of machine learning, recall, sorting, CV, NLP and other fields of algorithm models 3. Familiar with any of the following open source tools: Tensorflow, Pytorch, etc. 4. Have excellent learning ability and teamwork spirit 5. Have experience in upgrading and iterating industrial-grade search/recommendation/advertising systems 6. Those with excellent rankings in Tianchi and other machine learning competitions and ACM competitions are preferred 7. Priority will be given to those who have published papers in top international conferences.
