Системный дизайн инфраструктуры сбора данных для ИИ на примере рекомендательных систем

Speaker: Alexander Rumyantsev - Software Engineer, Social Media Practice of Integration and System Architecture Design: https://clck.ru/3Rd2FN This talk will demonstrate how to lay a solid foundation for AI products. We'll explore what happens before model training and why data quality determines 90% of success. You'll receive ready-made diagrams for designing a Data Collection system, which you can immediately apply to your projects. Using a real-world example, we'll explore: What a recommender system consists of and where data is "hidden" within it How to design a collection system so it doesn't break under load Why data quality monitoring is just as important as service monitoring A real-world case: how to bypass restrictions with high traffic volumes 1:36 Terminology check 4:52 ML project stages 9:00 Recommender system architecture 21:35 Questions and discussion