
Beyond the Model: Building AI Agent Harnesses That Actually Work
About this event
Part of Zürich AI Meetup | Secure Orchestration: this workshop takes place before the main meetup, which starts at 17:30 at the same venue. Main meetup: https://www.meetup.com/zurich-ai-meetup/events/316103725/
Two systems can use the same model and behave like radically different agents. The difference is the harness: the orchestration, state, tooling, guardrails, and deterministic scaffolding around the model.
In this 60-minute workshop, João Carvalho from Uthereal (https://uthereal.ai/) will cover:
• The puzzle: same model, radically different agent
• What is a harness
• State of the art and framework examples
• The economics: scaffold analysis and model agnosticity
• Build a minimal harness together (time permitting)
• Uthereal production demo
• Discussion: what should remain deterministic?
Registration requires approval on Luma: https://luma.com/r2i0trj3
Source: meetup