Innovathens

AI can write your Terraform, generate your CI/CD pipelines, and spin up infrastructure in seconds. So do we still need platform teams, DevOps practices, and carefully designed cloud architectures?

Your CEO doesn’t want SQL, neither does he or she want dashboards. What if instead you could give them an application that talks to your data? Join me to learn how Cloudwalker built Echo — a conversational analytics engine built on Amazon Bedrock and Amazon Redshift for the betting industry. I’ll share what worked, what broke, and what we’d do differently — from schema-aware prompting to guardrails that keep AI honest.

This talk will take you on a journey through the pursuit of performance in AWS serverless services. For those who have embarked on a similar path, you’d agree that it’s rarely a smooth ride. Complexities often arise, especially when navigating the mysteries of AWS (serverless) black boxes.

We’ll dive into XM’s backend architecture, which powers the real-time delivery of hundreds of thousands of financial market data to thousands of clients.

Amazon Aurora DSQL is a serverless distributed SQL database with virtually unlimited scale, highest availability, and zero infrastructure management. In this talk we’ll use a sample serverless application that uses Amazon API Gateway, AWS Lambda and Aurora DSQL to explore the architecture and features of this database. We’ll explore the scaling behaviour of DSQL, measure performance including cold starts, deep dive into its pricing and current limitations.

We’ll also use Strands Agent with Amazon Aurora DSQL to talk to our application using natural language. We’ll wrap up with some suggestions when currently to use DSQL comparing to DynamoDB, RDS and Aurora (Serverless) and what to pay attention to when migrating your existing application to Aurora DSQL.

Serverless computing abstracts infrastructure away, but that abstraction hides critical architectural decisions.

Let’s explore how to use Kiro Specs to make those decisions explicit, AWS MCP servers to understand what AWS services offer and what your architecture requires, and AWS CDK to generate infrastructure from specifications.

A live demo shows the complete flow: from spec to generated infrastructure.

At Skroutz we have lately been experimenting with building customised Agentic workflows for various security use cases.

This is when we started exploring Amazon Bedrock AgentCore and its ability to deploy server less agentic components easily as AWS-native primitives.

This session will cover our lessons learned, things that worked and things that didn’t so hopefully they can help the community experiment with Amazon’s latest agentic capabilities.

In the AWS world, some architectures shine, others… offer lessons on what to avoid. We’ll analyze real-world examples of anti-patterns: unmanageable architectures, ineffective scaling, and questionable technology choices. We’ll see what went wrong, turning every facepalm into a lesson. With humor and technique, we’ll learn to design robust and scalable solutions. If you want to learn from (others’!) mistakes while improving your cloud architectures, this is the talk for you!