Meitu Internet Technology Salon Session 7: Practices in Recommendation Algorithms, Big Data, and Personalized Recommendation

On December 17, Meitu Internet Technology Salon held its seventh session in Xiamen, livestreamed on Meipai, attracting over a hundred technical experts.

The salon focused on hot topics such as recommendation algorithms and big data. Speakers included senior engineers and researchers from Meitu and Hong Kong Hang Seng Management College.

Yu ChuQian – “Understanding Meitu’s In‑house Data Collection System Arachnia” described the need for reliable, high‑throughput log collection, the design of Arachnia, and how it addresses data reliability, latency, and resource usage.

Yang YaQiang – “Meitu Distributed Bitmap Practice” introduced bitmap technology for massive data computation, explained why existing open‑source solutions are insufficient, and presented Meitu’s own distributed bitmap service Naix.

Cai QiSen – “Practices and Exploration of Meitu Personalized Recommendation” covered the architecture of Meitu’s recommendation pipeline, challenges such as timeliness, cold‑start, and diversity, and shared optimization strategies.

Dr. Yang HaiQin – “New Perspectives on Improving Recommendation Technology” examined limitations of traditional collaborative filtering and proposed models based on data‑missing theory to boost efficiency and performance.

The event also highlighted Meitu’s large‑scale data processing infrastructure, the role of big‑data techniques in supporting recommendation systems, and offered a valuable learning platform for senior developers.