项目介绍

2026年9月30日

About the position

We invite applications for a Postdoctoral Research Fellow position in action-oriented modeling for trustworthy AI, available at the Department of Informatics, University of Oslo.

The starting date is no later than 01.03.2027 and as soon as possible after the appointment.

The appointment is a full-time position for a period of three years.

The fellowship is funded in full by TRUST for the whole period. No teaching, supervision, or research-assistance duties are required in the position, and the fellowship period can therefore not be extended to four years on that basis.

No one can be appointed for more than one Postdoctoral Research Fellowship at the University of Oslo.

The place of work is the Department of Informatics at the University of Oslo. The successful candidate will be part of the daily research and social life of TRUST – the Norwegian Centre for Trustworthy AI, and of the information systems research group.

Project description

The position is funded by TRUST, the Norwegian Centre for Trustworthy AI, which is a research centre dedicated to developing trustworthy AI systems that are accurate, fair, transparent, safe, and well-governed. The centre brings together researchers from computational and data science, law, social sciences, and philosophy to develop new AI technologies and study their societal impact in areas such as healthcare, industry, governance, security, and climate. TRUST is led by University of Oslo, SINTEF, and Norwegian Computing Center. With a consortium of over 70 partners from academia, industry, government, and civil society in Norway and internationally, TRUST’s research is theoretical, methodological, legal, and empirical in nature.

The postdoc will conduct the research project Action-Oriented Modeling for Trustworthy AI in Practice. The project addresses a persistent problem in applied AI: statistically accurate models routinely fail when they meet the organizations and ecosystems meant to use them. Models are misaligned with local procedures and decision processes, too rigid for resource-constrained operational systems, and opaque to the decision-makers who must act based on them. We treat this as a methodological problem rather than a technical one.

The primary deliverable of this position is a theoretical and methodological template that we call Action-Oriented Modeling. Rather than delivering finished algorithms, the template specifies how to build deliberately context-agnostic model architectures that expose methodological choices as parameters, and how local actors then configure, tune, and operate them. Its novelty lies in integrating three dimensions normally treated separately: configurable architectures that let domain experts shape model behavior without machine learning expertise; iterative field validation inside operational workflows; and governance mechanisms that transfer model ownership to local users.

The successful candidate will lead the development of this template into a generic, transferable, and theoretically grounded contribution. Your main tasks are to:

• Develop, formalize, and publish the Action-Oriented Modeling template. This is the central output
• Validate the template empirically through an action research design approach supplemented by participatory design methods
• Define governance principles for model ownership that let stakeholders revise models as operational contexts change
• Coordinate collaboration between the core development team, implementation partners, and local stakeholders, and act as a competence hub for action-oriented AI across TRUST

Two contrasting cases provide the empirical ground. The primary case is climate-sensitive disease prediction supporting public health decisions in low- and middle-income countries, built on DHIS2 and the action research network of the Health Information Systems Programme (HISP), with national health teams in ten countries. The complementary case is KoSy, the DHIS2-based municipal surveillance platform run by the Norwegian Institute of Public Health. We seek a candidate motivated by the methodological contribution to these and other empirical settings.

The main purpose of a postdoctoral fellowship is to provide candidates with enhanced skills to pursue a scientific top position within or beyond academia. To promote a strategic career path, all postdoctoral research fellows are required to submit a professional development plan no later than one month after commencement of the postdoctoral period.