Research Scientist - Large Language Models

About the role

The Institute of Foundation Models (IFM) is a research lab focused on building, understanding, using and risk-managing foundation models. As a Research Scientist with a focus on data-centric large language model (LLM) development, you will advance how LLMs reason, retrieve and interact with external information. You will work alongside researchers, data scientists and engineers on the core of foundation model training.

Key responsibilities

• Lead research and implementation of reasoning-enhanced LLM capabilities through novel data collection, architecture design and system integration.
• Design and implement pipelines to collect, curate and structure open-source and web-scale data for reasoning tasks, ensuring scalability and reproducibility.
• Build software supporting fine-tuning, evaluation and deployment of LLMs that interact with structured and unstructured knowledge bases.
• Collaborate with ML researchers on information retrieval, agentic search and RAG pipelines.
• Rapidly prototype tools, APIs and infrastructure for reasoning over external information, and build datasets to identify and analyse LLM failure modes.
• Communicate research findings internally and externally, including at top-tier conferences such as ACL, ICLR and NeurIPS.
• Take part in code reviews and engineering best practices, and represent MBZUAI at conferences and forums.

Required profile

• Master's in Computer Science, Data Science or a related field, or equivalent practical experience.
• PhD or equivalent research experience in Machine Learning, NLP or Data Science with a focus on reasoning and LLMs preferred.
• Experience with large language models, including fine-tuning, prompt engineering and multi-modal interaction.
• Preferred: experience designing agentic LLM systems, reasoning benchmarks or RAG pipelines, and a background in knowledge retrieval systems such as knowledge graphs and semantic search.
• Preferred: strong publication record at AI conferences and familiarity with production performance constraints.

Required skills

• Strong Python development skills, research-grade code and scalable data pipelines.
• Familiarity with collecting and processing large-scale datasets from open-source and web resources.
• Ability to work with ML infrastructure (model evaluation, optimization, debugging).
• Proactive approach to research questions and effective collaboration in cross-functional teams.