Job Opening at MBZUAI: Postdoctoral Researcher
Alham Fikri Aji
Department: Natural Language Processing / MBZUAI
Position Overview
The successful candidate will lead research initiatives in designing and developing comprehensive datasets for both training and evaluation of LLMs, with a focus on creating novel benchmarks and datasets that capture the complexity of real-world language understanding and generation tasks for Arabic and its dialects.
This role involves fundamental research in evaluation methodologies, corpus creation, and practical applications to assess state-of-the-art language models across diverse domains and capabilities. The position will involve close collaboration with linguists and domain experts.
Key Responsibilities
Dataset and Resource Development:
Design and support construction of datasets of dialectal Arabic NLP that can be utilized for training and improving LLMs.
Design and construction of innovative dialectal Arabic NLP benchmarks that evaluate previously unexplored aspects of language model capabilities.
Create robust annotation protocols and quality assurance frameworks for large-scale dataset construction.
Evaluation and Assessment:
Establish gold-standard evaluation resources and methodologies for Arabic LLMs.
Develop comprehensive evaluation frameworks that assess model performance across different Arabic dialects and domains.
Conduct systematic evaluation of existing language models on Arabic language tasks.
Research Dissemination:
Publish research findings in top-tier NLP conferences and journals
Release open-source evaluation tools and datasets to benefit the broader research community
Present work at international conferences and contribute to the Arabic NLP research community
Required Qualification
Education and Experience:
Ph.D. in Computer Science, Computational Linguistics, Natural Language Processing, Machine Learning, or closely related field
Demonstrated expertise in multilingual and/or Arabic NLP research
Demonstrated expertise in resource building and evaluation of NLP models
Strong track record of publications in top-tier NLP/AI venues (ACL, EMNLP, NAACL, LREC, NeurIPS, etc.)
Technical Skills:
Extensive hands-on experience with large language model evaluation and analysis
Proficiency in designing and implementing evaluation frameworks and metrics
Expert-level programming skills in Python and deep learning frameworks (e.g. PyTorch)
Experience with dataset construction, annotation, and quality control processes
Language and Communication:
Native or near-native proficiency in Arabic is strongly preferred
Excellent written and verbal communication skills in English
Demonstrated ability to work effectively in collaborative, interdisciplinary teams
Experience presenting research to both technical and non-technical audiences
Preferred Qualifications
Experience with multiple Arabic dialects and their linguistic characteristics
Background in corpus linguistics or language documentation
Previous experience developing datasets or evaluation benchmarks for underrepresented languages
Familiarity with dialectal Arabic computational challenges and solutions
Experience working with linguist collaborators and managing annotation projects
Cover Letter: Describe your research experience, interest in the position, and how your background aligns with our research goals
Curriculum Vitae: Complete academic CV including publications, projects, and relevant experience. Please highlight any prior datasets, benchmarks, or evaluation tools you have developed
Please send the application to alham...@mbzuai.ac.ae