3rd Youzan Big Data Technology Salon: Apache Kylin4, Data Governance, and AI Applications

On December 18th, the 3rd Youzan Big Data Technology Salon was successfully held. Originally planned for both online and offline formats, the event was adjusted to online-only livestream due to pandemic restrictions, yet still attracted over 200 participants who engaged with experts from Youzan, Kwai, and GeTui.

The salon opened with remarks from Jiang Bo, head of Youzan's Data Platform, who shared insights from merchant visits over the past year. He emphasized that data governance has evolved from a technical concern to a critical business issue. For example, merchants transitioning from distribution to direct sales need properly aligned product codes and attributes to enable business model transformation.

The event featured four technical presentations:

1. Apache Kylin4 Application at Youzan - Presented by Zheng Shengjun, Data Platform Lead at Youzan and Apache Kylin Committer. He discussed the capabilities and performance improvements brought by Kylin4.0 to Youzan's infrastructure, including upgrades that positioned Youzan as a leader in the OLAP field.

2. GeTui Data Intelligence Five-Step Method - Presented by Wang Yulun, Senior Data Development Engineer at GeTui. He explained how to practically apply complex machine learning techniques, including neural networks, reinforcement learning, and deep learning, by breaking down the process into five manageable steps.

3. Kwai's Sustainable Governance Based on Evaluation System - Presented by Sun Wei, Big Data R&D Expert at Kwai. He addressed challenges in data governance such as quantifying problems, identifying priorities, and maintaining motivation, showcasing how Kwai implements sustainable data governance practices.

4. Youzan Intelligent Copy Algorithm and Application - Presented by Kang Hongyu, Algorithm Engineer at Youzan. This session covered NLP applications including automatic translation, chatbots, sentiment analysis, and specifically how natural language algorithms automatically generate intelligent marketing copy.

The salon highlighted that as labor costs rise and data scales grow, efficient data insights and growth discovery become increasingly urgent. The shift from automation (efficiency improvement) to intelligent discovery (uncovering unknown insights from data) represents the current focus of data-driven businesses.