About Me
I am an AIGC technical leader at Kuaishou, where I work on commercial AdsLLM and AI Agents. My recent work focuses on multimodal generation, large language models, generative recommendation, advertisement video creation, and efficient deployment of large-scale AIGC systems.
I received my master’s degree from Peking University, China, under the supervision of Prof. Yadong Mu. My research interests span computer vision, multimodal learning, large language models, AI agents, AIGC, and recommendation systems.
Before that, I received my bachelor’s degree in computer science from Wuhan University, where I graduated with a GPA of 3.83/4.0 and ranked 1/38 in my major.
News
Four recent papers were accepted by CVPR 2026 and KDD 2026.
Adaptive Video Distillation: Mitigating Oversaturation and Temporal Collapse in Few-Step Generation
Narrative Weaver: Towards Controllable Long-Range Visual Consistency with Multi-Modal Conditioning
One paper was accepted by CVPR 2023.
Learning Instance-Level Representation for Large-Scale Multi-Modal Pretraining in E-commerce
Research Area
- Computer Vision
- Multimodal Learning
- Large Language Models and AI Agents
- AIGC and Generative Recommendation
Experiences
AIGC Technical Leader, Commercialization Algorithm Department, Kuaishou. (2025.01 - Present)
- Leading the development of AdsLLM and AI Agents for Kuaishou’s commercialization scenarios, including zero-material advertisement generation, intelligent customer service, and ad delivery agents.
- Built a zero-to-one advertising agent that can automatically generate complete AIGC short video advertisements from product images or landing pages, reducing production cost by orders of magnitude and scaling to hundreds of thousands of videos per day.
- Led the AI novel agent system from zero to one, covering LLM-based classification, structured understanding, highlight extraction, content generation, and quality evaluation for large-scale online deployment.
- Drove generative recommendation and personalized AIGC ad generation systems for large-scale advertising, and led research on efficient video generation, controllable long-range visual consistency, and multimodal advertisement editing.
Senior AIGC Algorithm Engineer, Commercialization GenAI, ByteDance. (2023.03 - 2025.01)
- Led AdsLLM for commercialization, covering continued pretraining, alignment, evaluation, and platform capabilities for advertising creation, moderation, customer service, and diagnosis.
- Built data cleaning, SFT construction, and quality evaluation pipelines for more than 40 business scenarios, and improved long-context, function-calling, RAG, and instruction-following capabilities.
- Established a comprehensive evaluation system combining automatic, GPT-based, and human review for both general and vertical-domain commercial tasks.
Computer Vision Algorithm Engineer, AI-Lab, ByteDance. (2021.10 - 2023.03)
- Led the video-to-product retrieval project for commercialization, enabling large-scale matching from videos to in-platform and external products with multimodal understanding.
- Built end-to-end capabilities including intent recognition, detection and tracking, product recognition, multimodal retrieval, and ranking, and supported multiple e-commerce and advertising scenarios.
Computer Vision Algorithm Engineer, AI-Lab, ByteDance. (2021.06 - 2021.10)
- Built multimodal models for advertisement understanding by combining visual and textual signals to improve CTR and CVR prediction in advertising systems.
Algorithm Engineer Intern, Alibaba Group. (2020.07 - 2020.10)
- Worked on advertising creative material mining and video understanding for Alibaba’s creative center, improving the diversity and scale of reusable creative assets.
Research Intern, Medical AI Lab, Tencent. (2018.07 - 2019.07)
- Worked on the core reading-comprehension model for the MedQA system and on automatic chest X-ray diagnosis and report generation.
- Engineering Intern, Tencent Cloud. (2017.07 - 2017.09)
- Responsible for the Panshi system decoupling and reconstruction. Separated the system’s web access layer from its data access layer; completed the modification of related data interfaces.
- Developed and maintained the related functional plugins of the system.
- Exchange Student in Hong Kong Baptist University. (2017.05 - 2017.06)
- Researched on blockchain technology.
Contact
Email: yongzhili at pku.edu.cn
Wechat: lyz327067530
Adress: Someplace, Beijing, China .
Welcome
