Workflows, Agents, and Everything In Between

Hack Session

About the session

There are a lot of ways to build an AI agent. A single smart prompt, a pipeline of LLM calls, a supervisor with specialist workers, a fully autonomous system. Each one has its place. Knowing which fits your problem is a different skill entirely. This session is about building that intuition.
 
We take one real use case and build five different agent architectures on top of it, then look at what actually happens. For each pattern we walk through the code, dig into the results, and figure out what it does well and where it falls apart. The goal is not to pick a winner. It is to understand the tradeoffs well enough to make a deliberate choice.

Speaker

Download Brochure