ArchitectureArchitecture
Conference50min
INTERMEDIATE

Event-Driven at Scale: Evolving Architectures with Kafka, Java, and AWS

This talk explores XM's transition to an event-driven architecture using Kafka, Java, Spring, and AWS. It discusses tool choices like Spring Kafka, Kafka Streams, Flink, Lambdas, and Kafka Connect, highlighting successes, failures, and integration strategies. Expect practical insights and lessons from scaling event-driven systems in a real-world fintech environment.

Spyridon Angelopoulos
Spyridon AngelopoulosXM

talkDetail.whenAndWhere

Thursday, June 12, 17:25-18:15
Room 1
talks.description
What does it really take to scale event-driven architecture in a real-world platform? Inthis talk, we’ll share how XM, a global fintech company for online trading, evolved fromsimple service-to-service calls to a sprawling, event-driven ecosystem built on Kafka,Java, Spring, and AWS.As our systems (and headaches) grew, we tackled each new use case with a differentcombination of tools, from Spring Kafka and Kafka Streams to Flink, Lambdas, andKafka Connect. We’ll walk through what drove each choice, what paid off, whatbackfired, and how we wired everything together into something that actually runs inproduction.Expect a practical, experience-based look at building and operating event-drivensystems at scale. You’ll see real-world topologies, lessons we learned the hard way, andthe kind of tips we wish someone had told us earlier.
event-driven
fintech
kafka
scaling
talks.speakers
Spyridon Angelopoulos

Spyridon Angelopoulos

XM

Greece

Spyros Angelopoulos is an Expert Architect at XM, working on the design and evolution of large-scale backend systems in the fintech space.

With over a decade of experience in Java, Spring, and cloud-native development, he has designed and delivered resilient, event-driven platforms that power critical trading infrastructure.

He is hands-on with microservices, Kafka, Kubernetes, and AWS — and occasionally even writes code that works on the first try.

He focuses on building scalable systems that hold up in real-world production environments.
comments.title

comments.speakerNotEnabledComments