在 Dcard 我們如何用 Python 打造推薦系統

Abstract

At our Dcard ML team, we rely on Python to create our recommendation system. In addition to the data science and modeling stages, we also use Python to build the program responsible for handling requests in real-world usage. To ensure our program's robustness, we incorporate various features of the Python ecosystem, such as abstract classes, in multiple areas. Today, we are excited to share how we utilize these features at different stages of the ML lifecycle. We'll also discuss any challenges we encountered and how we overcame them, as well as any best practices we learned along the way.

Description