AWS User Group Berlin Session - October 2026
Dear Community,
Another month with two great speakers gladly joining us. The agents are the focus, and we have great engineers on the lineup to tell us more. More details below, we look forward meeting you all at NLND Berlin!
18:30 - Warming up and networking chat
18:45 - Intro by AWS UG Berlin & NLND
19:00 - 19:45 - Tetiana Senna // Trust, but Verify: Observability and Evaluations for AWS Agents
Getting an AI agent to work in a demo is easy. Knowing whether it still works in production - after the next prompt change, the next model upgrade, the next edge case - is the hard part. Unit tests don't cover non-deterministic behavior, and manually reviewing a handful of runs doesn't scale.
This talk walks through the two practices that close that gap: observability (understanding what actually happened inside your agent) and evaluation (deciding whether the outcome was any good).
We'll dig into Amazon Bedrock AgentCore's Observability and Evaluations capabilities, compare the proven evaluation approaches - built-in LLM-as-a-Judge evaluators for automatic quality monitoring, and custom evaluators for business-specific correctness checks - with a candid look at what each one is good for, where it falls short. We'll also take a close look at AgentCore Optimization, AWS's newly previewed feature in this space.
19:45 - 20:05 - Networking break with food, snacks and drinks
20:05 - 20:50 - Milad Rezaeighale // Engineering AI Agents for Production
Building an AI agent has never been easier, but engineering one for production remains a discipline of its own. Frameworks and managed services make it simple to get a working prototype, yet the hardest problems only appear once an agent takes on real responsibilities.
In this talk, I share best practices for shipping agents to production, across authentication, harnessing, agent steering, context engineering, memory, and evaluation. These are the areas where design decisions have the biggest impact on an agent's security, reliability and cost. Drawing on real implementations built with Amazon Bedrock AgentCore and Strands Agents, the session focuses on the architectural patterns and trade-offs behind each of them.
Very Important: There is no "waitlist" for our regular sessions. However, there are limited seats available. If you want to make sure you can attend:
Register yourself with your "full name" here at meetup.com
Arrive on time - seats are first come, first serve.
As soon as there are seats available, you are welcome to join with your registration. In case there are no more seats available, we won't be able to let you join us this time.We thank you very much for your understanding!
Additional Information
This event is wheelchair friendly. Help us spread the word, and invite your friends & colleagues! If you're attending with wheelchair and need assistance, please mail us: organisers@berlinawsug.de for further details.
Would you like to host AWS UG Berlin events at your company? Register here
Would you like to speak at AWS UG Berlin sessions? Submit your talk here