This is a 40-hour, hands-on week on building and running AI directly on devices, without depending on the cloud. It’s aimed at professors, students, and professionals in engineering, networking, and computing who know some Python — no prior AI experience needed.
Participants move from ML fundamentals through computer vision, audio, and motion, training and deploying their own models on microcontrollers (such as the Arduino NICLA), and close the week by running a small language model on an edge device (Arduino UNO Q) and presenting a team project that integrates what they built. For Latin America and the Caribbean, this matters because edge AI works where connectivity is unreliable, budgets are tight, and data shouldn’t leave the site — and because the instructors and the capacity stay in the region.
Sponsored by the EDGE AI FOUNDATION and coordinated by Marcelo Rovai (UNIFEI, Brazil) with instructors from Colombia, Argentina and Ecuador, at Universidad Técnica del Norte in Ibarra, November 16–20, 2026, with 30 places available.