Capstone Project

The capstone is where everything you have learned comes together. Instead of one more isolated exercise, you will build a complete AI product end to end: framing the problem, collecting and cleaning data, choosing and training (or prompting) a model, evaluating it honestly, and deploying it so a real user can interact with it. You pick ONE of six projects and own it from idea to working demo.

Your six options: (1) AI Language Learning Platform — generate bilingual lessons and quizzes, like an Amharic-English tutor app. (2) AI Agriculture Assistant — answer crop, weather and pest questions for farmers. (3) AI Customer Support Agent — a chatbot that resolves tickets using a company knowledge base. (4) AI Medical Information Assistant — explains general health information with strong safety guardrails. (5) AI Educational Tutor — adaptive practice and feedback for students. (6) AI Business Analytics Platform — turns sales data into plain-language insights and forecasts.

For example, a developer in Addis Ababa building the AI Agriculture Assistant might start small: collect 200 common questions from teff and maize farmers, connect an LLM with a retrieval step over a local agronomy guide, and deploy it as a Telegram bot so farmers can ask in Amharic by phone. A narrow, well-tested project beats a broad, half-working one.