Across Africa, universities are racing to introduce new degrees and tracks in AI and machine learning. Behind the exciting announcements, many departments are quietly struggling with the same questions: should we standardise on Python, how do we connect an introductory programming course to serious AI work, and how can we design syllabi and projects that reflect African contexts rather than copy‑pasted curricula? Drawing on AI and data‑science courses and curriculum experiments I’ve run in East African universities, this talk presents a practical, Python‑first framework for structuring AI/ML programs in resource‑constrained settings. We will look at three pillars: foundational Python for data work taught with notebooks and core libraries, integrated math‑plus‑code modules where statistics and linear algebra are learned through Python, and a progressive AI track that moves from classical ML to capstone projects using local datasets and walks through an example course map. This talk is for lecturers, curriculum designers, community organisers, and advanced students who are shaping AI/ML learning paths. Attendees will leave with a simple checklist and concrete steps to identify gaps in their current offerings, align on Python as a common language, and draft a realistic learning path from a first Python script to a first AI capstone project.