Postdoc: LLMs for AI-Enhanced Educational Support

Vrije Universiteit Amsterdam; Amsterdam; Hybrid; EUR 41,000 - 64,000; Full time

Benefits offered by this job

Hybrid working

ABP pension

232 hours holiday

Job summary

Vrije Universiteit Amsterdam is seeking a Postdoctoral Researcher to advance LLM components supporting education professionals in developing and refining OPPs. You will explore prompt strategies, structured inputs, and potential fine-tuning to generate targeted OPP sections and quality feedback.

Collaborating with software developers and educators, you will evaluate outputs against a defined quality rubric and help integrate AI components into the Presentis system for user-centered evaluation.

Qualifications

• PhD in AI/CS/NLP or closely related field.
• Hands-on experience with Large Language Models and modern NLP methods.
• Excellent communication in English; Dutch preferred.
• Ability to work in a multidisciplinary setting with educators and software developers.
• Strong programming skills in Python; experience with ML/NLP frameworks.

Responsibilities

• Select and benchmark SOTA LLMs for Dutch-language OPP tasks.
• Develop and evaluate LLM components for generating OPP sections.
• Design prompting strategies and structured input representations.
• Investigate fine-tuning and retrieval-augmented generation value.
• Create rubric-based methods for detecting quality issues in OPP text.

Skills

LLM basics

Python programming

PyTorch

Hugging Face Transformers

Prompt engineering

Model evaluation

Multidisciplinary collaboration

English communication

Scientific writing

Education

PhD in AI/CS/NLP or related

Tools

PyTorch

Hugging Face Transformers

Job description

Vrije Universiteit Amsterdam is seeking a Postdoctoral Researcher to advance LLM components supporting education professionals in developing and refining OPPs. You will explore prompt strategies, structured inputs, and potential fine-tuning to generate targeted OPP sections and quality feedback.

Collaborating with software developers and educators, you will evaluate outputs against a defined quality rubric and help integrate AI components into the Presentis system for user-centered evaluation.