Presentation: How to Build Real-time Image Recognition with Apache Kafka

Talk:

This talk presents common problems associated with building production-ready real-time applications at scale. Solutions presented in the talk will be later demonstrated on a real-world use case – an image recognition project. The project uses traffic cameras in London to collect video feeds from different parts of the city. The pipeline relies on Apache Kafka and microservices running in Kubernetes. It extracts frames from the video stream and uses machine learning to recognise what is depicted in the video. The end outcome of this pipeline is displayed in real-time in a web application on the map.