Sergius Justus C. Nyah
AI Researcher & Software Engineer
About
I'm a practicing Christian and disciple of the Lord Jesus Christ, seeking and serving the Lord in the domain of His call upon my life as a Youth Pastor, Christian Writer, and Software Engineer at Christian Missionary Fellowship International (CMFI), a Global ministry founded by Zacharias Tanee Fomum in 1975. I previously served as a Youth Pastor at CMFI Molyko in Buea, Cameroon, where I also led the Music ministry. I'm quite passionate about leveraging technology to advance the Kingdom of God and empower the next generation of Christian leaders.
I am also a fresh graduate in Computer Science from University of Buea, with research interests at the intersection of NLP, multilingual AI, LLM alignment, and Geometric Deep Learning. I was selected as one of 32 global fellows for the MIT Summer Geometry Initiative — with only two fellows selected from Africa — through which I went on to work as a Research Fellow under Professor Justin Solomon's Geometric Data Group at MIT CSAIL, contributing to research on 3D cortical mesh segmentation using deep learning methods for neuroimaging, mentored by Dr. Karthik Gopinath from Harvard Medical School. I also worked on multilingual LLM evaluation and bias research at the Fatima Institute for Global AI Research in collaboration with researchers from the Stanford NLP Group and Georgia Tech.
Currently, I am an incoming research intern at EPFL (École Polytechnique Fédérale de Lausanne) in the LiGHT Lab under Professor Mary-Anne Hartley's AI and Healthcare team.
Beyond research and engineering, I enjoy music ministry and gospel instrumentation. I'm been a pianist since childhood, and picked up the bass guitar in Secondary school, and I'm not letting it go!. When I can, I support FC Barcelona, Current Laliga Champions, Força Barça!❤️💙
Featured Research
To Lie or Not to Lie? Investigating the Biased Spread of Global Lies by LLMs
LLMs spread misinformation unevenly — with significantly higher propagation in lower-resource languages and regions with lower Human Development Index scores. We introduce GlobalLies, a multilingual benchmark spanning 195 countries, 8 languages, and 6,867 entities — exposing systematic gaps in AI safety across global contexts.
Z. Khan · M. Dogan · I. Okoh · P. Sadeghi · S. Shrestha · S. J. Nyah · M. O. Mokhiamar · M. J. Ryan · T. Naous