Postdoctoral Researcher University of British Columbia
Abdellah EL MEKKI, Ph.D.
I work on natural language processing and inclusive large language models for linguistically and culturally diverse communities.

About
Research across languages, cultures, and modalities
I’m a Postdoctoral Researcher at the University of British Columbia (UBC), working on inclusive large language models adapted to specific languages and cultures across modalities (text, audio, and images), including low-resource and under-represented languages. I also work on research and applied science projects where machine learning intersects with other fields. My work has applied machine learning to network intrusion detection, biology (peptide sequencing), time-series modeling, anomaly detection, and a range of tabular-data problems. Across these domains, I have explored customized input-representation methods, from classical ML methods to adapted Transformer architectures. I have 10+ peer-reviewed publications in venues such as ACL, EMNLP, and NAACL. I hold a PhD in Computer Science from Mohammed VI Polytechnic University (Morocco) and previously completed a postdoctoral year at MBZUAI (UAE). Beyond my research, I have served as an Area Chair for ACL Rolling Review, as a reviewer for leading AI/ML venues and journals including NeurIPS, and as an organizer of workshops and shared tasks at top-tier AI conferences.
Currently Open to collaborations and consulting.
Research
Areas of interest
- Multilingual and cross-lingual NLP
- Multicultural NLP
- Low-resource and under-represented languages
- Large Language Models and instruction tuning
- Sequence labeling (NER, POS), morphology, and syntax
- Dialect identification and language variety modeling
- Domain adaptation and transfer learning
Recognition
Selected awards
Best Resource PaperACL 2025
Area Chair Award (Machine Translation and Multilingualism)IJCNLP-AACL 2023 · Official announcement
Publications
Selected work
NileChat: Towards Linguistically Diverse and Culturally Aware LLMs for Local Communities
Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing: EMNLP 2025
ProMap: Effective Bilingual Lexicon Induction via Language Model Prompting
Proceedings of the 13th International Joint Conference on Natural Language Processing and the 3rd Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics
Outstanding Paper Award
Domain Adaptation for Arabic Cross-Domain and Cross-Dialect Sentiment Analysis from Contextualized Word Embedding
Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies
Writing
Recent research notes
Alexandria: A Dialectal Arabic Machine Translation Dataset for Real-World Arabic MT
Alexandria is a human-translated Dialectal Arabic machine translation dataset for English-Arabic MT, low-resource machine translation, Arabic dialect benchmarking, and LLM evaluation across 13 Arab countries.
WVS2Persona: World Values Survey Wave 7 Personas for Culture-Aware AI
WVS2Persona is a Hugging Face dataset that turns World Values Survey Wave 7 respondent records into textual personas for culture-aware AI, persona-based prompting, and cultural alignment research.
Unsupervised Machine Translation in the Age of LLMs
Unsupervised machine translation still matters in the LLM era. This post explains how self-mined in-context examples can improve translation for low-resource languages without parallel data.