Vision-Language AI Scientist: Multimodal Foundations

Mohamed bin Zayed University of Artificial Intelligence; Sunnyvale (CA); On-site; USD 150,000 - 450,000; Full time

Benefits offered by this job

Comprehensive benefits

Bonus

401K Plan

Generous paid time off

Job summary

Mohamed bin Zayed University of Artificial Intelligence is seeking a Research Scientist on the Vision Language Model (VLM) team to advance multimodal foundation models that combine vision and language understanding with agentic capabilities.

You will work on model architectures, data recipes for pre‑training and post‑training, and evaluation benchmarks, applying large‑scale data processing, distributed training, and reinforcement learning methods for multimodal reasoning and agents.

Qualifications

• PhD or equivalent research experience in ML, CV, NLP, or Multimodal AI.
• Strong Python and PyTorch development skills for large‑scale ML research.
• Experience with distributed training systems and large‑scale model optimization.

Responsibilities

• Research and development of next‑generation Vision Language Models across pre‑training, instruction tuning, reasoning, and agents.
• Develop architectures and training methods for visual understanding, language reasoning, and tool‑use.
• Build and evaluate large‑scale multimodal datasets and benchmarks for VLMs.
• Collaborate across research and engineering teams; contribute to reports and open‑source software.

Skills

Python

PyTorch

Vision‑Language Models

Multimodal AI

Distributed training

Education

PhD or equivalent research experience

Tools

PyTorch Distributed

Megatron

Triton

CUDA

Job description

Mohamed bin Zayed University of Artificial Intelligence is seeking a Research Scientist on the Vision Language Model (VLM) team to advance multimodal foundation models that combine vision and language understanding with agentic capabilities.

You will work on model architectures, data recipes for pre‑training and post‑training, and evaluation benchmarks, applying large‑scale data processing, distributed training, and reinforcement learning methods for multimodal reasoning and agents.