Wesley Kambale is a Machine Learning Engineer, Software Engineer, and Google Developer Expert in Artificial Intelligence and Machine Learning. At the AI for Uganda Initiative, he directs foundational research in low-resource African natural language processing and leads a national team of artificial intelligence trainers who have trained more than 250 developers across Uganda.
His work focuses on closing language and cultural alignment gaps in East African languages through sovereign foundation models, machine translation, speech recognition, and open data infrastructure, including the curation of the AI for Uganda Luganda Prose Corpus and the AI for Uganda Luganda Question-and-Answer Pairs Corpus.
Beyond language technology, Wesley builds applied computer vision and predictive analytics systems for healthcare, agriculture, environmental conservation, and civic governance across Uganda.
Foundation language models, speech pipelines, and applied AI systems built for Ugandan communities.
Native Luganda and cross-lingual English causal language models engineered from the ground up to support conversational dialogue, analytical reasoning, translation, and cultural proverb interpretation across 12 Ugandan societal domains.
A custom transformer model trained from scratch for English-to-Luganda machine translation, establishing a strong regional benchmark for low-resource indigenous language translation.
An automatic speech recognition and transcription pipeline tailored for spoken Luganda that reduces Word Error Rate by 22 percent compared to baseline models on Ugandan speech evaluations.
Architected the MusawoAI radiological scan analysis platform, a mango pathology diagnostic system, a 300-site environmental conservation monitoring platform, and The Uganda Wall civic tracking platform.
Sovereign language corpora, research preprints, technical white papers, and textbooks.
Contains 32.3 million words of verified news articles, encyclopedic passages, literary paragraphs, and explanatory prose across 12 Ugandan societal domains.
Contains 9.49 million words of multi-turn conversational dialogue, bidirectional translation, proverb interpretation, and analytical reasoning traces across 12 Ugandan societal domains.
Co-authored foundational white paper establishing the architectural roadmap for sovereign, culturally grounded, and responsible artificial intelligence systems in Uganda.
Authored by W. Kambale. A practical textbook for African universities and software developers bridging foundational Python programming, data science workflows, and applied artificial intelligence.
Co-authored by W. Kambale and R. Sankara. Presents custom subword tokenization, curriculum pre-training, instruction fine-tuning, and grammatical agreement benchmarks for Luganda.
Includes 24,973 contemporary Luganda-English sentence pairs covering governance, public health, and education, alongside 30,844 aligned verses normalized for Bantu morphological pre-training.