I'm a Computer Science student at Uganda Christian University and Equitech Futures Applied Data Institute Fellow. My research focuses on building AI systems that work for African contexts—from NLP tools for low-resource languages to data infrastructure for public utilities. Currently developing analytics platforms for water utilities across Uganda, Cameroon, Malawi, and Lesotho, and investigating responsible AI deployment through research on spurious correlations in medical imaging models. I bring a multilingual perspective (French, Swahili, English) and experience bridging technical solutions with real-world infrastructure challenges.
Edge AI and NLP for low-resource contexts, with applications in public infrastructure and information access. Interested in how data systems can improve decision-making for essential services in resource-constrained environments, and in building ML tools that account for the realities of African data ecosystems.
Python, C/C++, SQL, JavaScript, HTML/CSS, React
PyTorch, TensorFlow, Hugging Face, scikit-learn, NumPy, pandas
AWS, Git/GitHub, Streamlit, FastAPI, LaTeX
Arduino, Raspberry Pi, Microcontrollers
Applied Data Institute
Building data solutions for water utilities. Deep Learning Indaba, Rhodes Forum, GandaBERT launch.
AIESEC Uganda
Led 46% membership growth to 22 active members. Managed budget and partnerships.
AIESEC / Uganda Industrial Research Institute
Represented AIESEC Uganda internationally. Embedded systems development at UIRI (May-Jul).
AIESEC (Remote)
Information Management role supporting global operations.
Uganda Christian University
Focus: AI, Data Structures, Embedded Systems, Cyber Security.
Luganda, spoken by millions, lacks basic NLP infrastructure for information verification
End-to-end pipeline from dataset creation to deployed web application. Created novel 2,600-article dataset, fine-tuned Multilingual BERT achieving 85.7% accuracy across 5 categories.
Water utilities across 4 African countries lack unified analytics for operational decisions
Production dashboard with AI-powered natural language querying. Role-based authentication, JMP service ladder visualizations, Google Gemini integration for context-aware data assistant.
Medical imaging models can learn spurious correlations that reduce reliability in deployment
Investigating shortcut learning in malaria detection models using Ghana health data. Focus on model reliability and responsible AI for global health applications.
Full-stack application with Django REST backend and React frontend. User authentication, CRUD operations, real-time status tracking.
Adaptive four-way traffic system using 8 ultrasonic sensors. Real-time vehicle detection with emergency vehicle override capability.
Motion detection and alert system with PCB design.
Coming Soon
Writing about NLP for low-resource languages, responsible AI, and building tech for social impact.
Have a question or want to work together? Reach out below.