About the position

About the Institute of Foundation Models

We are a dedicated research lab for building, understanding, using,
and risk-managing foundation models. Our mandate is to advance
research, nurture the next generation of AI builders, and drive
transformative contributions to a knowledge-driven economy.

As part of our team, you’ll have the opportunity to work on the
core of cutting-edge foundation model training, alongside
world-class researchers, data scientists, and engineers, tackling
the most fundamental and impactful challenges in AI
development. You will participate in the development of
groundbreaking AI solutions that have the potential to reshape
entire industries. Strategic and innovative problem-solving skills
will be instrumental in establishing MBZUAI as a global hub for
high-performance computing in deep learning, driving impactful
discoveries that inspire the next generation of AI pioneers.

Position Summary

As a Research Scientist in the Vision Language Model (VLM) team,
your role will be central to advancing state-of-the-art multimodal
foundation models that integrate visual understanding, reasoning,
and agentic capabilities. You will work on the research and
development of large-scale VLM systems, spanning model
architectures, data recipes for pre-training and post-training, and
evaluation benchmarks. The role combines cutting-edge research with
practical engineering, emphasizing large-scale data processing,
filtering, and weighting pipelines, distributed training systems,
and reinforcement learning algorithms and systems for multimodal
reasoning and agent development.

Key Responsibilities

Research and development of next-generation Vision Language

Models across pre-training, instruction tuning, reasoning, and
agents.

Develop novel architectures and training methodologies for

integrating visual understanding, language reasoning, and tool-use
capabilities.

Research efficient multimodal learning techniques, including

data-efficient training, long-context modeling, model modularity,
and inference optimization.

Build and improve large-scale multimodal datasets, synthetic

data generation pipelines, and evaluation benchmarks for VLM
capabilities.

Investigate multimodal reasoning, agentic behavior, OCR,

grounding, document understanding, chart understanding, and visual
question answering capabilities.

Contribute to technical reports, research publications, and

open-source software.

Represent MBZUAI at research conferences and industry events,

showcasing advancements in multimodal foundation models and
large-scale AI systems.

Mentor junior researchers and collaborate across teams to drive

impactful research initiatives.

Academic Qualifications

PhD or equivalent research experience in Machine Learning, Computer
Vision, Natural Language Processing, or Multimodal AI.

$150,000 - $450,000 a year

Salary Range

The posted salary range represents the company’s good faith
estimate of the compensation for this position upon hire. The
actual compensation offered may vary within this range depending on
individual qualifications, including but not limited to relevant
skills, experience, education, certifications, geographic location,
and specific business needs.

Professional Experience Minimum

Experience working with large language models and/or

vision-language models, including pre-training, fine-tuning,
evaluation, or inference.

Strong Python and PyTorch development skills for large-scale

machine learning research.

Experience with distributed training systems and large-scale

model optimization.

Familiarity with multimodal datasets and data processing

pipelines involving images, text, and video.

Understanding of modern deep learning architectures, including

Transformers, attention mechanisms, and multimodal fusion
techniques.

Experience with ML infrastructure, including model evaluation,

debugging, optimization, and large-scale experimentation.

Problem-solving and research skills with the ability to

independently drive research/engineering projects.

Effective communication and collaboration skills for working

across research and engineering teams.

Preferred Skills

Hands-on experience training or fine-tuning large Vision

Language Models or multimodal foundation models at scale.

Experience with distributed learning frameworks and

infrastructure such as PyTorch Distributed, Megatron, Triton, or
CUDA.

Research experience in multimodal reasoning, agentic systems,

tool use, OCR, grounding, document understanding, or multimodal
retrieval.

Experience with synthetic data generation, multimodal data

curation, or automated evaluation frameworks for VLMs.

Familiarity with efficient training and inference techniques

such as FlashAttention, quantization, tensor parallelism, pipeline
parallelism, or memory optimization.

Experience contributing to open-source ML software and

large-scale research codebases.

Strong publication record in leading AI conferences such as

NeurIPS, ICML, ICLR, CVPR, ICCV, ECCV, ACL, EMNLP, or related
venues.

Experience collaborating across research, infrastructure, and

product-oriented teams to deliver state-of-the-art multimodal
systems.

We may use artificial intelligence (AI) tools to support parts of
the hiring process, such as reviewing applications, analyzing
resumes, or assessing responses and identifying potential
inconsistencies or verification signals in application materials
based on available information. These tools assist our recruitment
team but do not replace human judgment. Final hiring decisions are
ultimately made by humans. If you would like more information about
how your data is processed, please contact us.