Room 4G05

In this workshop, participants will learn how to build a customized reasoning assistant powered by the Meta Llama models hosted locally on Ollama. The focus will be on developing a local, step-by-step reasoning tool that mimics complex thinking processes, enabling users to experiment with how an AI breaks down problems, explores multiple solutions, and reaches final answers.

Participants will dive into the fundamentals of deploying local models, interact with code designed to perform iterative reasoning, and understand how to fine-tune their system to meet specific requirements. We will walk through how to set up a working prototype, use best practices in prompting, and build interactive user interfaces using Streamlit.

By the end of this workshop, participants will leave with their own version of a step-by-step reasoning model, and practical experience that can be applied to other advanced AI-based tools and models.