Company
Date Published
Author
Daniel Jagielski, Gil Friedlis, Matt Mangia
Word count
3322
Language
English
Hacker News points
None

Summary

This webinar discusses how to build GenAI apps faster using Kafka Streams' real-time stream processing technology. The main goal is to detect compromised accounts and malicious user behavior by analyzing website login attempts. The analysis involves four types of analysis: device recognition, location recognition, botnet detection, and brute-force attack detection. The system uses Apache Kafka as the event streaming platform and Kafka Streams API for stateless and stateful processing operations. The solution is designed to be scalable, fault-tolerant, and efficient in terms of memory usage. The final design optimizes the pipeline by using a composite key for the join operator, eliminating the need for a stateful aggregate operator.