glosings0n/ZungumzaAI

Zungumza.AI bridges AI research and humanitarian response in Eastern DRC. We are engineering offline, multimodal AI models (ASR/TTS) for low-resource languages, starting with Kilendu. This technology will power interactive educational apps for displaced children, combating the digital divide and war trauma. # Zungumza.AI 🌍 **Zungumza.AI** bridges AI research and humanitarian response in Eastern DRC. We are engineering offline, multimodal Artificial Intelligence models (ASR/TTS) tailored for low-resource African languages, starting with Kilendu. Our mission is to power interactive, mother-tongue educational applications for displaced children in conflict zones, helping combat both the digital divide and learning loss. ## Table of Contents - About the Project - Key Features - Tech Stack - Project Status - Getting Started - How to Contribute - License - Contact ## About the Project Protracted conflicts in Eastern DRC have caused large-scale displacement. In many Internally Displaced Persons (IDP) camps, children have limited access to physical schools. At the same time, many EdTech systems and modern AI tools systematically exclude vernacular languages. Zungumza.AI addresses this gap by building a scientifically grounded linguistic corpus for Kilendu and developing edge-ready AI models. This enables educational applications to run fully offline on rugged, solar-powered tablets in emergency learning hubs. ## Key Features - Low-resource NLP: Custom Automatic Speech Recognition (ASR) and Text-to-Speech (TTS) for underrepresented dialects. - Edge-AI deployment: Optimized neural networks designed to run offline on low-end hardware. - Multimodal data collection: Tools and protocols for participatory field data gathering and linguistic annotation. - EdTech integration: APIs for connecting AI services to Flutter-based educational games and virtual tutors. ## Tech Stack - Machine Learning: TensorFlow/PyTorch, Hugging Face Transformers (for example wav2vec 2.0). - Application Development: Flutter and Dart. - Cloud and Infrastructure: Google Cloud Platform (GPU training), Firebase. - Data Annotation: Open-source linguistic annotation tools. ## Project Status This project is in the initial data-gathering and research phase. The repository will progressively include: …