I am currently employed at Sina Weibo, bringing 10 years of deep expertise in the field of internet algorithms. I currently serve as an Algorithm Team Lead, managing a team of five. I possess a mastery of deep learning algorithms and am proficient across various domains, including large-scale models, recommendation systems, natural language processing, computer vision, and video processing. I specialize in the practical implementation of algorithmic models within real-world business scenarios. I have spearheaded numerous core algorithm projects, leading my team to execute algorithm optimizations and model iterations that resulted in significant improvements across key performance metrics. I possess comprehensive team management experience, excelling in Agile development methodologies to cultivate highly collaborative and efficient algorithm teams that effectively balance technical depth with business value. Throughout my career to date, I have filed a total of 21 patent applications for algorithmic inventions.

The main achievements from the three most recent positions are:

• Led a project on a theme-based audio recommendation algorithm, resulting in the filing of three invention patents. Conducted a one-week A/B test (scaled to 50% of traffic), achieving a 12% increase in click PV, a 6.76% increase in click UV, a 5.51% increase in playback duration, a 0.52% increase in playback UV, and a 31.17% increase in playback VV. Following long-term optimization, the CTR rose from a monthly average of 1.3% in March to 1.9% in June—a year-over-year increase of 46%—thereby significantly enhancing the conversion efficiency of audio recommendations.
• Led a project focused on audio recommendation algorithms for categorized content, resulting in the filing of two invention patents. Conducted a one-week A/B test (scaled to 50% of the user base), achieving improvements of 9.91% in click PV, 6.85% in click UV, 5.61% in playback duration, 0.64% in playback UV, and 19.89% in playback VV. In the second half of the year, assumed responsibility for the audio project's KR metrics, shifting the primary quantitative recommendation metric from CTR to "Audio Playback Page Duration"; in Q4, the average duration on the audio playback page reached 97,000 minutes, achieving an 82% completion rate for the metric and successfully meeting the core business objectives.
• Led collaboration between the algorithms team and the Product, Engineering, and Business departments; drove the breakdown and implementation of algorithmic requirements; achieved deep integration between technology and business; and earned high recognition from business stakeholders.
• During my tenure at Yuanfudao, the English essay grading system, the multimodal "You Draw, AI Guesses" project, and the homework beautification project—all of which I spearheaded—were successfully deployed; additionally, 12 algorithm-related invention patents were secured during this period.
• During my tenure at Xiaomi, I achieved a groundbreaking breakthrough in mobile video super-resolution algorithms, shattering the technical barrier that previously restricted on-device super-resolution processing solely to hardware chips. My proprietary super-resolution algorithm achieved state-of-the-art performance on relevant benchmark datasets, and during this period, I secured six invention patents for these algorithms.