Multimodal Vision AI Engineer (LLM Integration)

Duncan & Ross Consulting; Abu Dhabi; On-site; AED 331,000 - 477,000; Full time

Job summary

Duncan & Ross Consulting seeks a Computer Vision Engineer to design, build, and deploy multimodal vision solutions that interpret and describe visual data for integrated AI applications.

You will collaborate with data scientists to fine-tune vision-language models, develop production pipelines, and deploy code using PyTorch, TensorFlow, and OpenCV while upholding ethical AI practices and data privacy.

Qualifications

• 3–7 years of experience in computer vision, deep learning, or multimodal AI.
• Strong proficiency in Python and frameworks such as PyTorch, TensorFlow, Keras, and OpenCV.
• Experience integrating LLMs (GPT, Claude, Gemini, or open-source models) with vision systems.
• Solid understanding of transformer architectures, CNNs, diffusion models, and attention mechanisms.
• Familiarity with multimodal datasets (COCO, Visual Genome) and evaluation metrics for vision tasks.
• Experience with cloud-based AI tools (Azure AI, AWS Sagemaker, Google Vertex AI).
• Ability to write clean, scalable, production-grade code.
• Strong analytical, problem-solving, and communication skills.

Responsibilities

• Develop and implement computer vision models for image classification, object detection, segmentation, facial recognition, and visual understanding.
• Integrate vision models with LLMs to build systems that interpret and describe visual content.
• Design AI pipelines that combine text, images, and video data for multimodal learning and reasoning.
• Utilize deep learning frameworks (TensorFlow, PyTorch, OpenCV) to prototype and deploy models.
• Collaborate with data scientists and AI researchers to fine-tune vision-language models for specific tasks such as visual QA, captioning, or scene analysis.
• Implement data preprocessing, augmentation, and annotation pipelines for large-scale image datasets.
• Conduct performance benchmarking, optimization, and deployment of models in production environments.
• Research and experiment with emerging techniques in Generative AI, multimodal transformers, and neural architecture optimization.
• Develop APIs and tools for internal teams to utilize vision + LLM capabilities.
• Ensure compliance with ethical AI practices, including bias mitigation and data privacy.

Skills

Python

PyTorch

TensorFlow

OpenCV

LLM integration

Multimodal AI

Transformer models

Cloud platforms

Education

Bachelor's or Master’s in CS/AI

PhD preferred

Tools

PyTorch

TensorFlow

Keras

OpenCV

ONNX

Job description

Duncan & Ross Consulting seeks a Computer Vision Engineer to design, build, and deploy multimodal vision solutions that interpret and describe visual data for integrated AI applications.

You will collaborate with data scientists to fine-tune vision-language models, develop production pipelines, and deploy code using PyTorch, TensorFlow, and OpenCV while upholding ethical AI practices and data privacy.