Ignatius EzeaniPER is a Research Fellow at the UCREL Research Centre, Lancaster UniversityORG and a Visiting Researcher at Microsoft Research Lab AfricaORG, NairobiLOC. His work brings rigorous NLP to IgboLANG, WelshLANG, and 21+ African and low-resource languages — from national corpora to population-scale deployment.
// annotation scheme borrowed from MasakhaNER — the first large-scale NER datasets for 21 African languages
// UCREL, Lancaster · MSR Africa, Nairobi
About
Of the world's roughly 7,000 languages, fewer than 100 are well served by current AI systems. Nigeria alone has over 500 — with Igbo, Hausa, and Yoruba spoken by a combined 150 million people. Dr Ezeani's research treats language technology as infrastructure, and starts from the conviction that its absence falls hardest on those already at the margins.
Over more than a decade at Lancaster University — extended through Microsoft Research Lab Africa — his work has spanned African language NLP, minority-language tools for Welsh in active government use, digital and spatial humanities, and real-world AI deployment at population scale. The unifying methodological commitment: NLP can be made to work rigorously and equitably in any linguistic or cultural context, provided researchers build the infrastructure from the ground up.
He holds a PhD in NLP (Sheffield, 2019), a 2024 CoSeC Fellowship from UKRI-STFC, and an OpenAI Researcher Access Grant, and serves as Associate Editor of ACM Transactions on Asian and Low-Resource Language Information Processing. He is a long-standing contributor to the Masakhane community and served on the Technical Advisory Panel of the Lacuna Fund's African language dataset grants.
Research at a glance
// syncing with Google Scholar…
Projects & grants
Principal researcher and lead developer. Architect of the LLM-based journey and affect extraction pipeline for historical testimony corpora — VHA Holocaust testimonies, the 1947 Partition Archive, and the TRC corpus. Pilot study produced 85,348 structured journey records from 892 testimonies.
Principal contributor to the first large-scale named entity recognition datasets and models for 21 African languages — now a foundational benchmark for African NLP.
Research Associate on the first large-scale contemporary Welsh corpus, building taggers and semantic annotation tools for a living minority language.
Lead developer of bilingual Welsh–English NLP tools deployed and in active use by the Welsh Government — a repeatable pathway from research to government-grade deployment.
Principal investigator. Established the evaluation benchmark for Igbo–English MT, in active use across the research community, alongside a Google-backed Igbo question-answering dataset within Masakhane.
Invited talks
Media & advocacy
Updates
Timeline embeds are sometimes blocked by browsers or by X itself — follow directly instead.
Follow on XTalks, project milestones, and community threads are shared on LinkedIn first.
View LinkedIn profile