📋 Job Overview

The role involves leading technical exploration and implementation of multimodal large models for image-text and video in WeChat Video Account's business scenarios like red dot push and local life recommendations, aiming to enhance content generation accuracy and user conversion efficiency. Responsibilities include optimizing post-training, domain fine-tuning, and generation control of multimodal models using business data to achieve creative copy generation, personalized recommendations, and integrated distribution strategies. The position also requires deep understanding of user behaviors and scene demands to fuse multimodal generation models with recommendation systems, improving real-time content delivery and relevance.