多模态大模型算法工程师-抖音电商
Core
Building and optimizing e-commerce domain multimodal large model foundations, integrating e-commerce knowledge to support business applications like video QA, cross-modal retrieval, and product classification.
Role type
Senior IC multimodal large model algorithm engineer (e-commerce)
Builds
E-commerce multimodal pre-training pipelines, image/video multimodal models, and related application chains for e-commerce products.
Domain
E-commerce, Computer Vision, Multimodal AI
Deliverable
production ML models
Required skills
Multimodal large model development, Computer Vision (image/video classification, detection, segmentation), Self-supervised learning, Model optimization, Algorithm abstraction, Research & publication
Preferred skills
Experience with Llava, Mini-Gemini, Qwen-VL, Shikra, Kaggle/COCO/ImageNet/ActivityNet competition awards, Top-tier conference papers (CVPR, ICCV, ECCV)
Technologies
Llava, Mini-Gemini, Qwen-VL, Shikra, LLM, Video/Text models
Responsibilities
Develop e-commerce multimodal pre-training links and models; Optimize self-supervised pre-training models for e-commerce images/videos; Abstract algorithm interfaces to improve model reusability; Share SOTA models and publish patents/papers to empower business units.
