Welcome speech by Michal and Lydia about the idea behind AWS Community Day Slovakia.

The most exciting part of your job, solving hard problems, is being reclaimed by AI agents that work 24 hours a day without fatigue.

Routine coding is no longer where your real value lives. The developers who lead this era move up the stack, think architecturally, and orchestrate intelligent systems.

This session shows you how AI is reshaping the development lifecycle, what the Renaissance Developer mindset looks like in practice,and how to position yourself as the architect of AI-driven solutions, not a passenger.

As organizations scale AI agent development, robust enterprise architecture patterns become essential. This session explores proven best practices for building production-ready AI agents using Amazon Bedrock AgentCore and its core components.

Learn how to work backwards from business problems and start with focused use cases. Discover practical strategies for secure runtime deployment, standardized tool integration, evaluation frameworks, and end-to-end monitoring.

Explore agents collaboration, shared memory, identity-based access controls, and integrated observability. Walk away with actionable strategies for scaling AI agents across your organization through proper governance and team structures.

Most cloud security teams have solved visibility. You can detect misconfigurations, flag drift, and generate findings all day long. But then what? The real problem was never seeing the issues, it was fixing them at scale. In this talk, We’ll share how we went from drowning in findings to setting up our own auto-remediation.

We’ll cover how to implement remediation workflows using AWS services like Security Hub, EventBridge, Lambda, and Config Rules, review open source alternatives worth considering, and explain why we ultimately chose full customization to handle our specific needs.

Whether you’re just starting to think about auto-remediation or already sitting on a backlog of findings you can’t act on, you’ll walk away with practical insights.

Generative AI prototypes are easy to build—but making them enterprise-ready is not. In this session, we will explore how LangFlow, a visual orchestration framework built on LangChain, can be integrated with AWS-native services to design scalable, secure, and production-grade GenAI workflows.

We will cover how to deploy LangFlow on AWS (ECS/Fargate or EKS), integrate Amazon Bedrock as the LLM provider, use Amazon S3 for document ingestion, and apply IAM, networking, and secrets management best practices. The session will also discuss architectural patterns for extensibility, observability, and governance—key concerns for real-world enterprise adoption.

Attendees will leave with a clear understanding of where LangFlow fits in the AWS GenAI ecosystem and how to move beyond PoCs toward production.

We all love Lambda - however, developers often overuse it.

In this session and true event-driven fashion, we will explore functionless architecture - connecting AWS services directly through native integrations rather than Lambda and SDK.

Eliminating the need to bundle and deploy these functions, as well as saving on cost. During the demo, you’ll see a functionless application build with AWS CDK, its direct integration to various AWS Services such as API Gateway, EventBridge, S3, DynamoDB, Step Functions, handling business logic and data flow.

Karpenter has changed how we manage Kubernetes nodes on EKS — but deploying it is just the beginning. The real savings come from how you architect your NodePools, split workloads between Spot and On-Demand instances, and configure consolidation policies.

In this session, we’ll explore Karpenter’s configuration capabilities in depth: NodePool strategies (e.g., 50/50 Spot/On-Demand split), instance type diversification, disruption budgets, and consolidation tuning.

A significant portion of the talk is a live demo where we’ll show how Karpenter works during scaling events (scaling to 3,000 pods), observe consolidation in real time, and demonstrate splitting workloads 50/50 across On-Demand and Spot instances.

You’ll leave with concrete configurations ready to deploy in your own clusters.

During my time at school, I was amazed by the buildings constructed by different civilizations, their similarities and, at the same time, their uniqueness.

Depending on whether it was a house, a farm, a fortress, etc., the planning varied, as it required not only its own design plans but also varied in location, the materials and machinery needed, the number and type of workers (builders, electricians, plumbers, specialists), among other things, and with all this we could then have an estimate of the time and cost of the work.

Does this sound familiar? In this session, I want to share my experience working with different types of applications. I have been fortunate (and old enough) to work with applications ranging from mainframe systems to new multi-agent use cases. We will see how to leverage services such as EC2, ECS, EKS, Lambda, and Bedrock AgentCore across different architectural types.

I’ll introduce AWS Nuke for systematic resource cleanup, a hard reset button for AWS! We’ll dive into exactly how to build a production-ready solution by elevating the standalone CLI tool into an automated service using a custom AWS Architecture.

You’ll learn how to enable your team, see the automation in action via a demo, and discover why you absolutely shouldn’t eat the brown M&M’s.

This talk will focus on how small-scale applications can effectively utilize AWS serverless technologies. We’ll discuss foundational services like Lambda, DynamoDB, S3 and CloudFront, demonstrating how to build modern web applications that start small and scale seamlessly.

Using frameworks like SvelteKit and runtimes like Node.js, attendees will learn about cost-effective design principles and architectural patterns that ensure their projects are sustainable and ready for growth. Discover how AWS can empower your side projects with minimal overhead, allowing you to focus on delivering value and innovation.

AI agents are easy to prototype. Running them securely, observably, and at scale? That’s where things get complicated.

In this session, we move beyond agent frameworks and focus on what it takes to operate AI agents in production using Amazon Bedrock AgentCore. While frameworks like Strands provide modular, composable agent architectures, AgentCore delivers the managed runtime, identity, memory, gateway, and observability services needed for real-world deployments.

We’ll explore:

• Secure agent runtimes with session isolation and VPC support
• Built-in identity and fine-grained authorization
• Memory services for short- and long-term context
• Gateway integration for MCP-compatible tools
• Native support for Agent-to-Agent (A2A) communication
• Observability and tracing with CloudWatch and OpenTelemetry

Through practical architecture examples, we’ll show how to take an open-source Strands agent and deploy it on AgentCore — moving from proof-of-concept to production-ready AI systems.

Search engines guide users from their initial goals to the information that fulfills those goals. The retrieved items are deemed relevant when they align with the user’s intent. Ensuring relevant results involves having the data in the catalog, analyzing it accurately, representing it through text and vectors, ranking it appropriately, and performing other actions.

In this session, discover how auction houses utilize vector search, powered by Amazon OpenSearch Service, to find similar items within their extensive database, enhancing the customer experience and preventing issues with counterfeit artworks, and see how it assists users in obtaining the most relevant results.

This session also uncovers some peculiarities in getting embeddings with AWS Bedrock and AWS SageMaker serverless inference for the abovementioned goals.

In global commodity trading, a single-day nomination delay costs thousands, yet estimating logistics for hundreds of kilotons is still a manual bottleneck.

We replaced this with a serverless pipeline using SageMaker and custom-trained model. To let users query data without writing a single line of code, we built a chatbot using Amazon Bedrock and Athena that converts natural language into SQL queries.

This session breaks down the architecture and deployment of these systems.

LLMs, AI agents, machine-learning models are all the rage now - everybody uses them.

Running them as Kubernetes apps in the cloud, however, is a bit tricky - at the very least, you need expensive specialized hardware (GPUs) and a mindset tweak to work around the special nature of these workloads.

Let’s delve into what you need to do, what tools are available, and what fun and frustrating challenges await you when you decide to run containerized GPU workloads in Kubernetes in the cloud today - might actually change tomorrow, who knows?

OpenPrime is a platform that allows non-experts to design and deploy production-ready AWS infrastructure using Infrastructure as Code.

Through a guided ClickOps experience and an AI assistant, users define what they need, while OpenPrime automatically generates best-practice, auditable IaC that follows proven architectural, security, and operational standards. The result is consistent, maintainable AWS infrastructure without requiring deep AWS or Terraform knowledge.

As you move into Senior+, Tech Lead, or Engineering Manager roles, your calendar fills up fast, and your time to actually build steadily shrinks. Yet your responsibility for technical outcomes, architectural decisions, and knowledge sharing doesn’t disappear.

This talk shows how an AI-Native Software Development Lifecycle, powered by Kiro, can become a real leverage for senior engineers and engineering leaders.

Through a hands-on story of vibecoding between meetings, we’ll explore a practical setup designed for the challenges that dominate senior roles: problem framing, architectural reasoning, documentation, and alignment. We’ll build a first MCP server / Kiro power and show how AI can amplify impact - helping us learn faster, reduce cognitive load, and keep architecture and documentation in sync with reality, even when hands-on time is scarce.

While the focus is on Senior+ and Engineering Manager workflows, the tools and patterns shown are just as valuable earlier in a career, offering a powerful boost for learning, exploration, and day-to-day development, and you’ll leave with concrete examples and a repeatable blueprint for applying an AI-native SDLC in your own context.

Food and drinks are two of my major hobbies, and I love mixing cocktails at home. But when I host friends, I always end up as the menu. People keep asking what I have and what I can make.

In this talk I will show how I built an AI bartender on AWS that lets guests interact with a drink menu using natural language. Instead of listing options or explaining what’s available, guests can ask for things like “something fruity” or “a gin drink”, and the AI Bartender respond with suggestions powered by Amazon Bedrock.

We will go deep into how I built the AI agent with Amazon Bedrock using Strands Agents, and AgentCore providing conversation memory, identity, and a gateway for tool access. I’ll show how streaming responses, with Amazon API Gateway, are implemented for a great chat experience. MCP tools play a important role providing the agent with data, like ingredients, and menu stored in Aurora DSQL. Finally how serverless and event-driven patterns are used to decouple AI processing from APIs.

This is a technical, demo-heavy session with code walkthroughs and architecture deep-dives. We will explore agentic AI with tool calling workflows, conversation memory, and event-driven AI integration. You’ll walk away with insights on how to build intelligent, context-aware AI assistants that work with real data on AWS.

Securing and managing network traffic efficiently is crucial for modern cloud architectures. AWS Cloud WAN, combined with Network Firewall and Route 53 DNS Firewall, provides a robust framework for protecting both egress and east-west traffic.

This presentation explores how these security services enhance resilience, prevent threats, and enforce compliance across distributed environments. By integrating firewall policies and DNS security, organizations can strengthen their cloud infrastructure against cyber threats while maintaining optimal performance.

Learn how to build a secure, scalable, and resilient network with AWS’s advanced security solutions.

What happens when photography domain expertise meets AWS service flexibility? This session demonstrates how to bridge specialized knowledge with cloud architecture, creating solutions that AWS services weren’t originally designed for. Using AWS Community Day DACH 2024’s 8,000-photo challenge, we’ll explore how Amazon Rekognition, Bedrock, and Step Functions became an unexpected photography automation pipeline.

Learn the methodology for translating industry-specific workflows into serverless architectures, discovering how AWS services can be creatively combined beyond their intended use cases. Attendees will gain frameworks for identifying automation opportunities in their domains and practical strategies for connecting disparate AWS services to solve problems they weren’t originally built to address.

How do we know that our terraform in Git is in sync with Deployments?
How do we ensure automatic Terraform deployments?

As Infrastructure as Code (IaC) adoption accelerates, maintaining synchronization between version-controlled Terraform configurations and deployed infrastructure has become increasingly critical.

This presentation explores the challenges of ensuring GitOps alignment and examines the most prominent solutions for automating Terraform workflows.

Whether you’re managing a startup’s infrastructure or orchestrating enterprise-scale deployments, this session will equip you with the knowledge to make informed decisions about automating your Terraform workflows while maintaining the critical link between code and deployed resources.

In 2025, Agentic AI has demonstrated its ability to address a wide range of business domains — from financial services to marketing and retail. As successful pilots and early deployments deliver tangible value, organizations are increasingly pushing agentic systems from experimentation into production environments.

While Agentic AI promises autonomy and advanced reasoning, it also introduces a new class of production risks that traditional software systems — and even earlier Generative AI applications — do not face. These risks often surface only after early success, when autonomous behavior, non-determinism, and system-level interactions begin to scale.

In this talk, we examine the most common failure modes observed in real-world agentic AI deployments and discuss how to address them through robustness-aware system design. We explore architectural patterns, guardrails, and observability layers that help teams move beyond fragile prototypes and build agentic systems that are reliable, governable, and production-ready on AWS.

End speech and thank you by Lydia and Michal