Senior Vision-Language AI Research Scientist

Lever, Inc.; Sunnyvale (CA); On-site; USD 150,000 - 450,000; Full time

Job summary

MBZUAI invites applications for a Research Scientist role in the Vision Language Model team to advance state-of-the-art multimodal foundation models. You will research architectures, data recipes, and benchmarks for large-scale VLM systems, integrating visual understanding, language reasoning, and agentive capabilities.

You will collaborate with world-class researchers and engineers, publish results, mentor junior researchers, and help scale training pipelines with reinforcement learning

Qualifications

• PhD or equivalent research experience in Machine Learning, Computer Vision, NLP, or Multimodal AI.
• Experience with large language models and vision-language models, including pre-training, fine-tuning, evaluation, or inference.
• 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 methodologies for integrating visual understanding, language reasoning, and tool-use capabilities.
• Research efficient multimodal learning techniques and data-efficient training.
• Build and improve large-scale multimodal datasets, data generation pipelines, and evaluation benchmarks for VLMs.

Skills

LLM/Multimodal

Python

PyTorch

Distributed training

Education

PhD or equivalent

Tools

Megatron

Triton

CUDA

PyTorch Distributed

Job description

MBZUAI invites applications for a Research Scientist role in the Vision Language Model team to advance state-of-the-art multimodal foundation models. You will research architectures, data recipes, and benchmarks for large-scale VLM systems, integrating visual understanding, language reasoning, and agentive capabilities.

You will collaborate with world-class researchers and engineers, publish results, mentor junior researchers, and help scale training pipelines with reinforcement learning