Museum Route Agent

A conversational agent that provides routes between different locations in Belo Horizonte, with a primary focus on the city’s various museums. Developed as Practical Assignment for the Introduction to Artificial Intelligence course at UFMG in the first semester of 2025.

Image: Agent Architecture

Code available at https://github.com/carlabferreira/MuseumRouteAgent

Objectives and Key Features:

• Acquisition and utilization of the Belo Horizonte city graph.
• Use of search algorithms for pathfinding.
• Local execution of the Agent, using the ollama_chat/qwen3:8b model and defined Tools.
• Rendering/Exporting the obtained route as a .png image

Image: Route between “Museu de História Natural da UFMG” and “Casa Kubitschek”

Technology Stack:

• Ollama Model (qwen3:8b)
• Python libraries (smolagents, networkx, matplotlib, osmnx and more)

Example

Image: Screenshot showing two route queries and the agent’s responses.