Project Director
- Ulrich Paquet
Project Manager & Workshop Lead
- Michael Alummoottil
NLP AI Expert
- Ayman Saeed
- Similoluwa Adetoyosi Okunowo
- Toky Iriana RAJAOFERA
- Maharavosoaniaina Mari-Mar RAKOTONIRAINY
Empowering Universities to Lead Africa’s AI Future
A programme designed to equip African university lecturers with the skills, tools, and curriculum to teach state-of-the-art AI.
AI Capacity Building For African Universities
Leveraging the Google DeepMind AI Research Foundations course and a blended learning toolkit developed by pedagogy experts at UCL in collaboration with African science educators, the programme is designed to deliver scalable, sustainable, and locally relevant instruction in AI and Large Language Models across African universities. Supported by Google.org and implemented by FATE Foundation and AIMS South Africa.
Explore The Curriculum
Why This Matters
Africa’s universities are producing the continent’s next generation of scientists, engineers, and innovators — but most lack the faculty expertise, up-to-date curriculum, and hands-on resources to teach advanced AI effectively. Students who want to build language models, train neural networks, or work on real AI research often have nowhere to turn locally.
AI Research Foundations changes that. Through a Train-the-Trainer model, we upskill your lecturers — turning them into certified “AI Champions” — so they can deliver a world-class, localized AI course to your students, at your university, in your context. The programme runs from 2026 to 2028, reaching 30 universities across Ghana, Kenya, Nigeria, and South Africa.
Learn More
The AI Research Foundations team at AIMS
Lorem ipsum dolor sit amet, consectetur adipiscing elit.
The Curriculum - 8 Technical Courses
Learning Objective: Establish the foundational concepts required to implement a small language model from scratch.
Learning Objective: Master tokenization, embeddings, and text data preparation
Learning Objective: Build and evaluate neural networks; diagnose overfitting and training issues
Learning Objective: Implement self-attention and multi-head attention mechanisms
Learning Objective: Apply transfer learning and parameter-efficient fine-tuning with LoRA
Learning Objective: Use RLHF to align model outputs for safety and helpfulness
Learning Objective: Optimise inference with quantization and GPU memory management
Learning Objective: Build a real-world AI application: dialogue system, text classification, or summarisation
