Research Scientist at Meta, working on Threads Social Media AI — tech lead on retrieval and early stage ranking. Previously Software Engineer at Google (2011–2025), building image agents for Display Ads and machine-learned spelling for Search.
LinkedIn GitHub Google Scholar
- 🔍 Ex-Google
- ⏳ 19 YOE
- 💡 15 AI/ML patents
- 📚 20+ academic publications
- ⚡ >1M TPU/GPU-hours of AI training
- 🤖 Top 0.5% Agent Author in Meta Agent Market
- 🏆 2025 Google Tech Impact Award — Image Agent for Ads
Expertise
AI Agents & Application LLM
- Image agent orchestration: multimodal workflows for image ad creation; RAG and embedding for output quality
- Mixed agent swarm: tooling, time/budget management, deadlock prevention, workspace isolation, red-team and critique
- Agent in ranking funnel: retrieval, list ranking, content/user understanding, agent teacher for distillation
AI for Recommender Systems
- Social media ranking: retrieval, early stage ranking, user interaction sequence modeling, generative recommender
- Display ads ranking: CTR/eCPM/CVR prediction and selection; ads format design with computer vision
- Search engine ranking: ML based keywords and query understanding
Work Experience
Research Scientist, Meta
CA, USA · Oct 2025 – now
- Threads Social Media AI: tech lead on retrieval and early stage ranking
- AI agent for ranking; agent for ML productivity
- Generative recommender: HSTU based dual/cross encoders, generative retrieval
Software Engineer, Google
CA, USA · Oct 2011 – Oct 2025
- Display Ads (2015–2025): image agents for automatically generating and enhancing image ads; full lifecycle of Asset Based Creative products
- Search (2011–2015): tech lead of “Showing-results-for” V2 and “Aggressive Spelling”
Postdoctoral Scholar,
GRASP Laboratory,
University of Pennsylvania
2009 – 2011
- Visual autonomous agents planning and tracking (ICCV 2011)
Postdoctoral Scholar, Department of Statistics,
UCLA
2007 – 2009
- Active basis object detection; trackability of video representations
Team Leader, Lotus Hill Institute
2006 – 2007
- Intelligent video surveillance system (MIT Technology Review; ICPR 2008; CVPR 2008)
Education
- 2003 – 2007 · PhD in Pattern Recognition and Intelligent Systems
National Laboratory of Pattern Recognition, Institute of Automation, Chinese Academy of Sciences. Advisor: Prof. Songde MA - 2000 – 2003 · Master of Technology in Vehicle Engineering, Wuhan University of Technology. Advisor: Prof. Wanfu YANG
- 1996 – 2000 · Bachelor of Technology in Vehicle Engineering, Wuhan University of Technology
Professional Activity
Reviewer for IEEE T-CSVT, IEEE T-IP, IEEE T-PAMI, ICCV, Journal of Electronic Imaging, Journal of Field Robotics, Machine Vision and Applications, Pattern Recognition, Pattern Recognition Letters, and the National Natural Science Foundation of China.