About this role
Section: Overview AIDHealth Inc., a Nigerian not-for-profit organization focused on healthcare access and humanitarian support, is seeking a Hausa NLP Data Scientist to develop, evaluate, and maintain language models for analysing Hausa-language content.
Section: Responsibilities
- Benchmark listening platforms to assess their coverage of relevant Hausa content.
- Fine-tune African-language models, such as AfroXLMR or AfriBERTa, using project-coded Hausa data.
- Automate coding categories only when models achieve a macro-F1 score of at least 0.80 against human coding.
- Monitor model accuracy monthly and retrain models when performance declines.
- Conduct time-series analysis and comparisons across study phases.
- Maintain the secure, encrypted data warehouse and project code repository.
Section: Requirements
- MSc or higher in Computer Science, NLP, Data Science, Statistics, or Computational Linguistics.
- At least 3 years of professional experience in applied NLP or machine learning.
- Hands-on experience fine-tuning transformer models.
- Experience with African or other low-resource languages, ideally Hausa.
- Strong Python skills, including Hugging Face, pandas, and scikit-learn, plus working knowledge of SQL.
- Native Hausa speaker with fluent English.
Section: Preferred Experience
- Publications or open-source contributions in Hausa NLP, such as AfriSenti or NaijaSenti.
- Experience with topic modelling (BERTopic) and time-series analysis in R or Python.
Section: Skills and Attributes
- Rigorous measurement and honest reporting of model accuracy.
- Ability to explain technical limitations to non-technical colleagues.