About

Hi! I am David Graus. I study how AI can make information more accessible, and I think the public debate about AI and algorithms is missing the point.

I’m an Assistant Professor at the ILLC, University of Amsterdam, where I lead the ICAI OpenGov Lab. Before returning to academia, I spent years in industry as a Lead Data Scientist at Randstad and FD Mediagroep, building the kinds of systems people worry about: algorithms that decide what news you see and who gets hired.

That experience taught me that AI brings more value, and the challenges of incorporating it well are more interesting, and solvable, than dystopian headlines suggest.

What I work on

My research sits at the intersection of information retrieval, NLP, and societal impact. What brings it all together: using AI for better information access, and to help people find, search and discover what matters: whether that’s the right job, relevant news, or government documents that should be public. Read more about my Research, or see a summary below:

Currently, I lead the ICAI OpenGov Lab, a research lab set up in partnership with the Organization for Information Management at the Government of the Netherlands, where we study AI for government transparency. We build tools like WooGLe (a search engine for Dutch government documents) and research how to bridge the information gap between government and citizens. I supervise two PhD students: Damiaan Reijnaers on explainable document representations and Maik Larooij on advanced search, plus several master students working on open government, AI, and law.

Previously, at Randstad (2020–2024), I led research on job recommender systems: high-stakes algorithms that influence careers and livelihoods. My team worked on fairness and bias mitigation, skills extraction, and career path recommendations. I co-organized the RecSys in HR Workshop from 2021–2024.

At FD Mediagroep (2018–2020), I led the AI team building a news recommender system for het Financieele Dagblad (FD), and on BNR SMART Radio, a personalized news radio app that won a Marconi Award and Dutch Interactive Award.

My PhD (2012–2017, supervised by Maarten de Rijke) was on semantic search for e-discovery: helping lawyers find relevant documents in massive collections. During my PhD, I spent time at the University of Maryland in Washington DC, and at a summer internship at Microsoft Research in Redmond, US, where I worked on analyzing Cortana logs, which led to a best paper at UMAP 2016 and a patent.

Algorithms aren’t evil

The public debate around AI is dominated by non-technical voices, and I think we’re getting it wrong. We fear filter bubbles, assume personalization will end democracy, and treat data collection as something sinister. I try to bring a different perspective, grounded in building these systems.

This doesn’t mean I’m naive about risks. I’ve worked on bias in hiring algorithms, consulted on digital forensics in high-profile court cases, and teach a course on AI safety, regulation, and fairness. But I think computer scientists need to be part of the conversation.

Some things I’m proud of in this space:

Beyond research

I’m involved in various extra-curricular activities beyond the lab:

  • Board treasurer at SETUP, a foundation for public engagement with technology
  • Member of the Commissie Persoonsgegevens Amsterdam, advising the city on data protection and algorithms
  • Public Engagement Fellow at Amsterdam AI, talking with citizens about AI in places like NEMO and public libraries
  • Publicity co-chair for ECIR 2026 (and, in what feels like another lifetime, ECIR 2014)

Background

I have an unusual path for an academic: I started in humanities (Media Studies BA, UvA) before switching to computer science (Media Technology MSc, Leiden). Between degrees, I worked as a science editor at NTR, writing and producing for radio and online. That media background still shapes how I communicate research.