Hi, I’m Daniel van Strien, a Machine Learning Librarian at Hugging Face.
I work on practical AI and the tools that make it easier to use. That includes building datasets, training and evaluating models, and contributing to the command-line tools, infrastructure and documentation that people and AI agents use to get things done.
I’m particularly interested in small, task-specific models, and how AI agents can help people build, evaluate and use them. A recurring question in my work is how to tell whether a model is actually useful for a particular task—not just whether it scores well on a benchmark.
Before Hugging Face, I worked in libraries and digital humanities, including on the British Library’s Living with Machines project. Documents and cultural collections remain a recurring focus—from evaluating optical character recognition (OCR) on historical scans to extracting structured information and making images easier to discover and reuse. That background also shapes my interest in building technology with the people who know the material and understand what would make it useful.
I share experiments, practical guides and findings on this blog. I’m also writing AI Design Patterns for Information Professionals, a book in progress. You can find examples of my work on the Selected work page.