About Me
I am a PhD candidate at the SAINTS Lab at the Department of Computer Science at the University of Copenhagen (UCPH), supervised by Raghavendra Selvan and Erik Dam and funded by the Independent Research Fund Denmark. My PhD work revolves around the sustainability of “hyperdatafication” in AI, which is a phenomenon describing the industrialised production and accumulation of data for developing AI models (see more in Wilson et al., FAccT, 2026). My work seeks to address this by identifying its costs, including environmental impacts, economic inequality, and social harms, specifically the working conditions experienced by data workers.
I am currently a visiting researcher at Samsung (October 2026 – March 2027), where I work on profiling resource consumption and hardware utilisation of agentic AI systems on edge devices.
I hold an MSc degree in Computational Physics from the Niels Bohr Institute, UCPH, where I explored physics-informed machine learning as a way to reduce energy use and carbon emissions of machine learning models. My master’s thesis was supervised by Raghavendra Selvan and Jens Hesselbjerg Christensen.
Keywords: Sustainable, resource-efficient, frugal, and responsible AI and data practices.
Upcoming Activities
2026-11 Giving a talk on AI - det koster på klimaet at the HK TechWeek 2026 on 17/11.
2026-11 Giving a joint talk with Emil Njor on Reducing Resource Costs of Frontier AI by Identifying and Managing AI-Waste at the Digital Tech Summit 2026 on 4/11.
2026-10 Giving a talk on Hyper-datafication and its Environmental, Social, and Economic Costs at the Brandenburgisches Zentrum für Medienwissenschaften, Potsdam, Germany, on 22/10.
Recent Activities
2026-10 Attended the AAAI/ACM Conference on AI, Ethics, and Society in Malmö, Sweden, sat on the Responsible AI Student Panel, and presented ongoing work on egocentric data collection and annotation.
2026-10 Attended D3A 4.0 in Nyborg, Denmark, and won the Jury’s Best Poster Award! Also, co-organised the Deep Dive Workshop on Resource-Aware & Constrained AI at D3A 2026.
2026-09 Presented a webinar on Assessing Sustainability of AI at the High Energy Physics Software Foundation Seminars.
2026-09 Interviewed by journalist Mette Holdbæk in the podcast Verden Forsøgt Forklaret about AI data workers, their working conditions, and hyper-datafication more broadly.
2026-09 Interviewed by Uniavisen: Forsker: »AI bliver aldrig en bæredygtig teknologi«.
2026-09 Presented a webinar on Bæredygtig AI: udfordringer og løsninger (in Danish) at the Center for AI, DMJX.
2026-08 Paper accepted to ECCV 2026, Limit Workshop! See preprint on arXiv: Performance-Carbon Trade-Offs across Architectural Biases in Shear Flow Forecasting.
2026-08 Attended the Eastern European Machine Learning Summer School 2026 in Cetinje, Montenegro, and won the Best Poster Award!
2026-06 Presented our paper How Hyper-Datafication Impacts the Sustainability Costs in Frontier AI at ACM FAccT 2026, Montréal, Canada.
2026-06 Panelist in the UCPH panel on Love, Climate Crises, and AI at Folkemødet 2026 (“The People’s Meeting”), Bornholm, Denmark.
2026-05 Speaker at the P1 Workshop on Green AI at the Pioneer Centre for AI.
2026-04 Paper accepted to ICML 2026! See preprint on arXiv: Stop Preaching and Start Practising Data Frugality for Responsible Development of AI.
2026-04 Invited speaker and panelist at Sustainable AI in Practice, hosted by the Lund Stem Cell Center, Lund University.
2026-03 Co-organiser and moderator of the SAINTS’26 workshop, a half-day workshop on Sustainable AI for Sciences.