Stanko Kuveljić

Stanko Kuveljić

Staff ML Engineer at SmartCat.io

Stanko Kuveljić is a Staff ML Engineer focused on building reliable AI systems under real-world constraints. He has led production deployments across e-commerce and regulated environments, designing architectures that separate probabilistic model reasoning from deterministic business logic. His current work centers on agent reliability, behavioral testing, and the engineering discipline required to make generative AI systems observable, testable, and trustworthy in production.

Agenda Talks


AI Con

16:50 to 17:25
04 Nov 2026

Caging the Agent: From Prompt Chaos to Deterministic State

Agent workflows look stable in a demo. They call the right tools, follow the happy path, and feel like magic. Then a real user shows up.