Cameroon AI Research Lab

Welcome to the official hub of LAB-237. We are an academic and community-driven research group dedicated to advancing Natural Language Processing (NLP), speech technology, and Artificial Intelligence resources for Cameroonian languages, West African Pidgin, and regional dialects.

🎯 Our Mission

• Language Preservation: Documenting and digitizing indigenous Cameroonian languages.
• Dataset Creation: Building high-quality, open-source benchmarks for low-resource languages.
• Model Adaptation: Fine-tuning small language models (SLMs) to make state-of-the-art AI efficient and accessible locally.
• Community Empowerment: Creating accessible AI tools to bridge the digital language divide across Central and West Africa.

📦 Key Projects

• Camspoken NLI: The first foundational Natural Language Inference dataset optimized for Cameroonian languages and Pidgin variants.
• Localized Text-to-SQL: Building benchmarks and models that translate Cameroonian languages and Pidgin queries into executable database scripts (SQL).

⚖️ Open Source Licensing

To maximize global collaboration, scientific progress, and commercial accessibility, the Cameroon AI Research Lab operates under an open-source framework:

• Unless stated otherwise, all software code, scripts, model architectures, and core assets published by this organization are licensed under the Apache License 2.0.
• Individual dataset file repositories may inherit this permissive Apache infrastructure or utilize data-specific permissive open licenses (like CC BY-NC 4.0) to ensure unrestricted community integration.

🤝 Collaborate

We welcome collaboration from researchers, lecturers, students, engineers, computational linguists, data scientists, and other contributors interested in building AI resources for Cameroon and related low-resource settings.

Individual contributors and creators are credited in the documentation of their respective projects.

Founder & Lead Architect: Tiani Pekins Ebika

Affiliation: Department of Computer Engineering, University of Buea, Cameroon.

Contact: tiani@tianipekins.com

Find our codebase, training scripts, and evaluation pipelines on GitHub.