Millicent Ochiengの肖像

Millicent Ochieng

Research Scientist

概要

I am a Research Scientist at Microsoft Research Africa and a PhD researcher in Computer Science at the University of the Witwatersrand (opens in new tab). My work centers on multilingual and multicultural AI, with a broader goal of building AI systems that are more linguistically representative, culturally aligned, and equitable.

At Microsoft Research Africa, I spearhead multilingual tokenization research, developing approaches that improve linguistic representation and downstream model efficiency across global languages. I also lead work on multicultural AI alignment, examining how multilingual and multimodal systems reflect cultural context and how community-grounded evaluation can surface misalignment across diverse underrepresented communities.

My PhD research focuses on cultural alignment in multilingual language models for African languages, under the supervision of Prof. Bruce Bassett (opens in new tab). I study how regional and cultural variation can be represented, modeled, and evaluated through region-conditioned data, LLM alignment methods, and pluralistic evaluation.

Prior to Microsoft Research, I completed an MSc in Data Science at the African Centre of Excellence in Data Science – University of Rwanda (opens in new tab), where my research focused on transfer learning for precision agriculture. I later conducted research at Harvard University, applying machine learning, computer vision, and statistical methods to sustainable agriculture.

I have published in leading NLP and AI venues including EMNLP, NAACL, LREC and contribute actively to the African NLP research community through Masakhane (opens in new tab) and AfricaNLP Workshops (opens in new tab) where I serve as Workshop Organizer and Communication Chair.

Beyond research, I serve as WiDS Kenya (opens in new tab) Lead Ambassador, leading a community of 400+ members advancing gender equity in AI and data science. I also serve on the WiDS Worldwide (opens in new tab) Ambassador Advisory Council, contributing to the broader global ambassador community.

🤝 Let’s connect

I’m always happy to connect around multilingual or multicultural AI, tokenization and multilingual model efficiency, cultural alignment and equitable AI, AI for low-resource and underrepresented languages, and collaborations across academia, research labs, and Global South initiatives.

News: