Vision-Language AI Research Scientist

Institute of Foundation Models; Sunnyvale (CA); On-site; USD 180,000 - 250,000; Full time

Job summary

Institute of Foundation Models in Sunnyvale, CA seeks a Research Scientist to advance state-of-the-art Vision Language Models and multimodal AI systems. You will explore architectures, data recipes, and training strategies for large-scale VLMs alongside researchers and engineers.

Responsibilities include data processing pipelines, distributed training, and evaluating multimodal reasoning and agentic capabilities.

Qualifications

• PhD or equivalent research experience in Machine Learning, Computer Vision, Natural Language Processing, or Multimodal AI.
• Experience with large language models and/or vision-language models across pre-training, fine-tuning, evaluation, or inference.
• Strong Python and PyTorch development skills for large-scale ML research.

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 visual understanding, language reasoning, and tool-use capabilities.
• Research efficient multimodal learning techniques, including data-efficient training and long-context modeling.
• Build and improve large-scale multimodal datasets, synthetic data generation pipelines, and evaluation benchmarks for VLM capabilities.
• Represent MBZUAI at research conferences and mentor junior researchers.

Skills

Python

PyTorch

Distributed training

Multimodal reasoning

English communication

Education

PhD or equivalent research experience in ML/CV/NLP/Multimodal AI

Tools

PyTorch Distributed

Megatron

Triton

CUDA

Job description

Institute of Foundation Models in Sunnyvale, CA seeks a Research Scientist to advance state-of-the-art Vision Language Models and multimodal AI systems. You will explore architectures, data recipes, and training strategies for large-scale VLMs alongside researchers and engineers.

Responsibilities include data processing pipelines, distributed training, and evaluating multimodal reasoning and agentic capabilities.