Dr. Markus Nissl

A smarter model narrows the gap between plausible output and correct output. It doesn't close it. When agentic systems fail in production, the failure usually lives in the engineering around the model — the context that didn't get passed, the evaluation tier that wasn't built, the observability gap that hid it.

This talk gives you a framework for engineering reliable agents — drawn from a newly published book on Agentic Engineering.

You'll learn:
- Why the bottleneck has shifted from generation to evaluation, and where the time actually goes now
- The four dimensions of context most teams confuse for one thing
- Why standard logging hides the most important agent failures
- The five-tier evaluation hierarchy that decides whether a production agent holds up
- The diagonal that separates safe autonomy from silent drift

Provider-agnostic, framework-agnostic. Principles that outlive whatever SDK wins next quarter.

Dr. Markus Nissl develops AI systems that connect reasoning with action — from intelligent agents and knowledge graphs to interactive human-AI interfaces. As an entrepreneur and researcher, he focuses on making artificial intelligence transparent, explainable, and truly usable. His work bridges technological depth with practical design, showing how AI can empower both developers and organizations to create systems that think, act, and collaborate.