Innovation in action

Explore the solutions and research developed by our team and the talented members of our student cohorts.

Our projects

Solutions developed by the AI for Uganda core team.

Screenshot of the Pearl model

Pearl model: English to Luganda translation

NLP PyTorch Flask Translation

The Pearl model (23M-translate-v2) is a Transformer-based NMT model for English-to-Luganda translation, built from scratch with 23M parameters. This version supports longer sequences (512 tokens), deeper architecture (6 layers), a 16K-token vocabulary, and uses the original Transformer learning rate schedule.

Hugging Face View Code

Student projects

Innovative capstone projects from our program graduates.

Screenshot of pneumonia detection

Pneumonia detection using chest x-rays

Computer Vision ResNet-50 Python

This app uses a ResNet50-based deep learning model to detect pneumonia in chest X-ray images. It applies transfer learning to deliver fast, accurate predictions that support early diagnosis and assist healthcare providers.

Live Demo View Code