Andrea is currently CTO at Radicalbit, Milan. His main works have been focused on streaming technologies, machine learning, and performance-boosting. Andrea co-authored «flink-jpmml» project; he loves to spread the voice about how to regulate machine learning end-to-end lifecycle and streaming applications. He co-authored «Benchmarking Data Flow Systems for Scalable Machine Learning» science paper at DIMA Group, TU Berlin.
Development of a Kafka-Powered Advanced Stream Commerce Platform
During the pandemic, GoLive — an Advanced Stream Commerce platform — is born to help its customers to find new ways to reach their clients. Leveraging the Kafka know-how gathered over years, Radicalbit built a scalable Live Streaming AI-powered platform, by which we deliver in one place real-time analytics and streaming prescription to our customers, so they can enable the Continuous Intelligence pattern through their organization. During the Video Live sessions, we put together in Kafka topics clickstreams coming from hundreds of thousands of mobile devices and business customers’ data, using Kafka connect. We process the streams and we apply them against our Deep Learning models to produce streaming prescriptions using Kafka Streams and our MLOps system. Finally, we deliver them to GoLive data stores to build up the analytics experience, achieving each aforementioned step with pure (fun) Kappa architecture. Since the GoLive platform itself is built on top of Kafka, we also will highlight the advantages of using the same streaming platform to achieve asynchronous communication between microservices and real-time web-socketing.