Floris Nzabakira

Automated extraction and cleaning of 5 million+ text data from different online sources for a targeted set of languages.
Researched, designed and trained state-of-the-art machine translation models, leveraging the structure of African languages and their low resource nature.
Designed and optimized machine translation models, reducing workload for the development team while assisting them in deploying the models.

Improved Rwandan citizens' knowledgeability by designing and implementing a machine learning chatbot named “Bakame” that aims to assist Rwandan citizens in accessing government services on the Irembo platform.
Championed training the team on how to create a chatbot using RASA to ensure that our chatbot could accurately understand and process human language.
Collaborated with cross-functional teams on machine learning projects and contributed to upcoming community projects' design and development.

Developed a cutting-edge facial recognition system utilizing Facenet technology to accurately mark attendance in real-time while recording the information in databases, ensuring the optimization of real-time video streaming.
Developed, tested, and dockerized RESTful APIs to access the ML models, data pipelines, and project-specific components, resulting in improved accessibility, scalability, stability, and reliability, and reducing downtime while increasing user satisfaction.