About 360 AI Research
360 AI Research is 360 Group’s core AI research team. Our mission is to advance world-class technology and create measurable value through real-world deployment.
We focus on understanding and processing multimodal information across images, text, and video. Our research spans computer vision, natural-language understanding, multimodal understanding, and multimodal generation. These technologies are deployed across 360 Group’s businesses—including internet content distribution, enterprise digitalization, and AIoT—supporting tens of millions of devices and serving hundreds of millions of users.
We turn frontier AI research into dependable productivity, use technical leadership to enable new business directions, foster an engineering-led professional culture, and share our practical experience with the broader research community.
Honors & Awards
National Key Laboratory for Big Data Engineering
Open Laboratory for Multimodal Big Data
Wu Wenjun AI Science and Technology Award
Our Team
Dawei Leng
Head of Multimodal Foundation Models
Dr. Dawei Leng earned his PhD from Tsinghua University. He is Deputy Director of 360 AI Research and a board member of the China Society of Image and Graphics. He leads the group’s multimodal foundation-model strategy and core research team, with a long-standing focus on computer vision, multimodal learning, and production deployment.
Chunyu Xie
Head of Multimodal Understanding
Chunyu Xie holds a master’s degree from Beihang University. His work focuses on cross-modal retrieval, vision-language learning, and open-world perception, with an emphasis on making algorithms dependable across real-world products.
Bo Cheng
Head of Multimodal Generation
Bo Cheng holds a master’s degree from Xi’an Jiaotong University. His research covers text-to-image and text-to-video generation as well as controllable generation and editing for professional creative workflows.
Zetao Huang
Head of Inference Optimization
Zetao Huang holds a master’s degree from Tianjin Normal University. He focuses on inference acceleration, model quantization and compression, and cloud-edge deployment that brings foundation-model capabilities to production efficiently and reliably.
Open Positions
Research Scientist — Multimodal Understanding
Responsibilities
Work on internet image-text search, recommendation and advertising, AIoT devices, and enterprise digitalization. Track, research, and deploy frontier methods in image-text understanding, cross-modal alignment, multimodal foundation models, open-world object detection, and GUI agents. Provide technical leadership and production support for business teams.
Requirements
- Required: strong understanding of cross-modal image-text alignment and CLIP, including its capabilities and limitations
- Preferred: hands-on experience training or fine-tuning CLIP or related models
- Preferred: hands-on experience training or fine-tuning self-supervised vision models
- Preferred: established research experience in image or video classification, detection, semantic segmentation, tracking, or grounding
- Preferred: practical R&D experience in open-world object detection
- Master’s degree or above; PhD preferred. Publications at leading CV, ML, or AI conferences are a plus