Multimodal AI Benchmarking Research Assistant

Keele University; United States; Teletrabalho; USD 56 000 - 74 000; Tempo integral

Resumo da oferta

Keele University is seeking a motivated team member to join the School of Computer Science and Mathematics to develop an open multimodal AI compliance and adversarial auditing benchmark for evaluating AI systems in regulated sectors.

The post holder will design and implement benchmark tasks across text, structured/telemetry, audio and vision-language data, and develop LLM evaluation pipelines including robustness testing.

Qualificações

• Master's degree in Artificial Intelligence, Machine Learning, Data Science, Computer Science, Cyber Security or a closely related area.
• Strong programming skills, particularly in Python, and practical experience with ML and/or LLMs/agentic AI.
• Experience with PyTorch, Hugging Face, multimodal AI, adversarial machine learning, LLM evaluation, LLM agents, cybersecurity, Git/GitHub or reproducible experimental pipelines would be desirable.

Responsabilidades

• Develop and run LLM evaluation pipelines and implement adversarial robustness testing.
• Prepare and process multimodal datasets across text, telemetry, audio and vision-language data.
• Evaluate open-weight LLMs using reproducible experiments and contribute to the evaluation framework.

Conhecimentos

Python programming

ML/LLMs experience

Communication skills

Formação académica

Master's degree in AI/ML/DS/CS/Cyber Security or related

Ferramentas

PyTorch

Hugging Face

Git/GitHub

Descrição da oferta de emprego

Keele University is seeking a motivated team member to join the School of Computer Science and Mathematics to develop an open multimodal AI compliance and adversarial auditing benchmark for evaluating AI systems in regulated sectors.

The post holder will design and implement benchmark tasks across text, structured/telemetry, audio and vision-language data, and develop LLM evaluation pipelines including robustness testing.