Kakao Community Data Center Recommendation Team Introduces Large-Scale Recommendation System Construction and Technical Culture

Kakao's Community Data Center manages everything from the collection to the provision of company-wide data, and within it, the Recommendation Team is responsible for recommendation technology, the culmination of data analysis. This team designs the experience of users discovering information through recommendation features in major services such as Melon, Media Daum, and KakaoTalk Gift.

Scope of Large-Scale Recommendation System Construction

The Recommendation Team performs the entire process of building recommendation systems, including data pipeline configuration, machine learning algorithm research, and large-scale real-time API operations. The team includes various crew members such as developers, data analysts, and planners, who collaborate horizontally from idea generation to service application. Additionally, the team aims for a culture of growth by sharing knowledge across fields such as platform, analysis, and planning through regular seminars and study groups.

Recommendation Technology Trends and Developer Competencies

In the field of recommendation systems, issues such as the limitations of traditional methodologies pointed out in the RecSys 2019 paper and solutions for reinforcement learning and feedback loops are actively discussed. Recommendation developers are required to have not only Python, C++, and GPU programming skills but also a comprehensive understanding of systems including OS and networking, as well as experience with large-scale data processing. Kakao plans to recruit talent with these competencies to create new data-driven user experiences.