Everyone wants a Three-Michelin-Star AI Agent. It looks so simple on the menu: a dash of LLM, a reduction of conversations, a garnish of autonomous tools, and a sprinkle of RAG. But in the kitchen of DIY development, "simple" is the first ingredient in a recipe for disaster.

Join Chef Szymon Kulec from RavenDB for a behind-the-scenes look at the "tasting menu" of agent construction. First, we'll inspect the complexity of our base ingredient: the LLM. Then, we'll try to save our fingers from the universal blender of MCP and tools. From there, we move to data visibility, the ultimate question of "Should everyone know how the sausage is made?", before tasting our RAGoût to see how we can make it cook faster. Have we mentioned that we'll reduce our conversation sauce? We will!

At the end, we’ll finish with a GenAI twist that might save you a fortune.

A talk for technical people, showing ingredients what is needed to create an agent.