Requirements
Candidates must possess a Ph.D. in Computer Science, Machine Learning, Artificial Intelligence, or a related field with a strong focus on NLP or LLMs. A strong publication record in leading AI/ML/NLP venues, hands-on experience in training/fine-tuning/evaluating LLMs, proficiency in deep learning frameworks (PyTorch/TensorFlow), solid understanding of LLM architectures and advanced training techniques, and strong Python programming skills are required. Excellent communication and collaboration skills are also essential.
Description
Job description / Role
Job Type
Full Time
Job Location
Abu Dhabi, UAE
Nationality
Any Nationality
Salary
Not Specified
Gender
Not Specified
Arabic Fluency
Not Specified
Job Function
IT - Software & Web Development
Company Industry
Software & Internet Services
Description
Mohamed bin Zayed University of Artificial Intelligence (MBZUAI) is the world’s first AI-focused university. Founded in 2019, MBZUAI is a research-oriented graduate-level university that has grown to host over 80 world-class faculty, over 330 graduate-level students, and is already ranked in the world’s top 25 by CSRankings for AI-related fields. Located in Abu Dhabi, the university aims to become a global leader in AI education and research by providing cutting-edge programs in artificial intelligence, empowering students to shape the future of technology. MBZUAI is committed to fostering innovation, collaboration, and ethical practices in AI to address global challenges.
About the role
This project is a collaboration between MBZUAI and a major industry partner to develop proprietary, domain-specific large language models (LLMs) for the oil and gas industry. These models will form the core of an agentic system designed to dramatically reduce decision-making time and enhance operations for end-users.
This position will focus on fundamentally improving the complex reasoning capabilities of LLMs. The postdoctoral associate will be responsible for researching, developing, and implementing advanced techniques to create models that can solve complex mathematical, physics, and reasoning problems specific to the energy sector. This is a unique opportunity to work at the intersection of foundational AI research and high-impact industrial application, under the supervision of Prof. Martin Takac and Prof. Salem Lahlou.
Key responsibilities
Design and implement advanced training and fine-tuning methodologies (Reinforcement Learning from Human Feedback, Reinforcement Fine Tuning, Mixture-of-Experts, etc.) to enhance the reasoning capabilities of LLMs.
Conduct domain adaptation of open-source models using proprietary data from the oil and gas sector.
Implement and refine retrieval-augmented generation (RAG) techniques and integrate knowledge graphs to improve model accuracy and contextual understanding.
Conduct rigorous evaluation of models on both domain-specific datasets (MCQ, QA) and standard industry benchmarks (MMLU, GPQA, MATH, GSM8K, etc.).
Ensure models are safe and helpful, preventing the generation of harmful content or sensitive information.
Stay up-to-date with the latest literature on LLM reasoning, developing and implementing novel ideas.
Pioneer innovative techniques and original ideas aimed at fundamentally improving the reasoning abilities of large language models, both for project-specific applications and for the wider research community.
Publish research findings in top-tier AI and NLP conferences and journals.
Work with and mentor graduate students involved in the project.
Qualifications
A Ph.D. in Computer Science, Machine Learning, Artificial Intelligence, or a related field with a strong focus on natural language processing or large language models.
Minimum:
A strong track record of relevant publications in leading AI, machine learning, or NLP conferences and journals.
Hands-on experience in training, fine-tuning, and evaluating large language models.
Proficiency in deep learning frameworks such as PyTorch or TensorFlow.
Solid understanding of LLM architectures and advanced training techniques (e.g., RLHF, RAG, Mixture-of-Experts).
Strong programming skills in Python.
Excellent communication and collaboration skills.
Preferred:
Demonstrated research experience in LLM reasoning, mathematical problem-solving, or agentic AI systems.
Familiarity with knowledge graphs, graph-augmented language models, or preference-based optimization methods.
Experience working on collaborative, multi-disciplinary research projects, especially with industry partners.
Experience mentoring junior researchers or students.
What we offer
Opportunity to work on a high-impact, industry-relevant research project with a major energy partner.
A stimulating research environment at MBZUAI with leading experts in AI.
Competitive compensation and benefits package aligned with top-tier academic markets.
Significant opportunities for professional development and travel to major conferences.
A one-year appointment with the possibility of extension based on performance and project needs.
Apply Now