Internship | AI‑Driven Dataset Curation: Leveraging VLMs to Understand and Improve Dataset Quality

About this internship

High‑quality datasets are the backbone of trustworthy Artificial Intelligence. Yet, strengthening AI systems through careful dataset curation remains an often underestimated part of model development and deployment. In this internship, you will explore how natural‑language‑driven image quality assessment can enhance dataset curation practices. Where traditional methods rely on numerical indicators such as contrast, noise levels, or resolution, modern multimodal models enable richer, grounded descriptions of image quality. Building on ideas from work like GROUNDING‑IQA: Grounding Multimodal Language Models for Image Quality Assessment, this project investigates how image‑level quality descriptions can be combined into dataset‑level profiles, revealing coverage gaps, biases, and systematic quality inconsistencies. The focus will be on real‑world datasets, with particular attention to how well these quality assessment approaches generalize to niche domains, long‑tail categories, or fine‑grained classes. The results will contribute to more structured, transparent, and explainable dataset curation workflows.

What will be your role?

You will design and implement a pipeline that generates and analyzes natural language image quality descriptions, perform dataset-level aggregation and keyword analysis, and evaluate how well these methods generalize across different datasets and domains. Through experiments, you will assess how language-based quality profiling can complement traditional numerical quality metrics in data validation workflows. The precise focus and methods can be tailored to match your interests and strengths.

During your internship, you will write a thesis documenting your research. If the results are promising, we may explore the possibility of a joint publication.

What we expect from you

You are in the final stages of your master’s degree in artificial intelligence, computer science, data science, or a related field. You have experience with Python, deep learning, and preferably some familiarity with vision-language models. You are curious about how AI can be made more interpretable and robust in real-world settings.

We expect you to work from our office in The Hague (Oude Waalsdorperweg 63) at least two days per week, but preferably more often. Your supervision team will consist of experienced researchers in computer vision and deep learning, and are actively involved in related projects. Furthermore, they regularly supervise students, often with overlapping topics that encourage collaboration and knowledge exchange.

What you'll get in return

You will be properly supervised during your work placement and be given the scope for you to get the best out of yourself. Furthermore, we provide:

• A highly professional, innovative internship environment, within a team of top experts.
• A suitable internship allowance (615 euro for wo-, hbo- and mbo-students, for a full-time internship).
• Possibility of eight hours of free leave per internship month (for a full-time internship).
• A free membership of Jong TNO, where you can meet other TNO professionals and join several activities, such as sports activities, (work-related) courses or the yearly ski-trip.
• Use of a laptop.
• An allowance for travel expenses in case you don’t receive an OV-card.

TNO as an employer

At TNO, we innovate for a healthier, safer and more sustainable life. And for a strong economy. There is all the space and support for your talent and ambition. You work with people who will challenge you: who inspire you and want to learn from you. Our state-of-the-art facilities are there to realize your vision. What you do at TNO matters: impact makes the difference.

pieter.piscaer@tno.nl