.png?locale=en)
Tame your AI-Agent - Evaluation and Observability
26.06.2026
|
13:00
-
14:00
h
The access link will only appear if you sign up for this event.
Language:
English
You're building AI agents with LangGraph or a similar framework - but how do you know they're actually working?
Unlike traditional software, LLM-powered systems are probabilistic. The same request can produce different outputs, and failures are often subtle, hard to reproduce, and difficult to diagnose. As agents become more autonomous, evaluation and observability become essential engineering practices.
In this session, I will show you how to bring rigor to AI agent development using an observability platform such as LangSmith.
You'll learn how to:
- Build evaluation datasets and measure agent performance systematically
- Move beyond "it seems to work" with quantitative evaluation metrics
- Use traces and observability tools to understand agent behavior step-by-step
- Identify and debug common failure modes in agent workflows
- Integrate evaluation into your development process to iterate with confidence
The session will include real-world examples from agent projects, showing how evaluation and observability can help you move from prototype to production-ready systems.
Who it's for
Technical decision makers, CTOs, AI engineers, and developers building LLM applications and agents.
Format
Live demos, code, traces, and practical examples.
Recommended for: CTOs, Product Managers, Product Leaders, Developers
Hosted by
.png?locale=en)
The Agent Native Product
I work with B2B software companies to rebuild their products AI-native.
Alex Key
alex.key@kawunu.comRecommended Events
Talk
Data Science & AI
Conference
Data Science & AI
Tame your AI-Agent - Evaluation and Observability
26.06.2026
|
13:00
-
14:00
h