We take on a small number of students and visitors whose questions overlap with the three research themes. There are no paid staff openings listed at the moment.
None at this time.
Agentic alignment
Design and evaluate lightweight monitors that flag off-policy actions in an agent stack, and measure false-positive cost against detection latency.
Application-layer privacy
Quantify how masked prompts degrade extraction, transformation, and reasoning tasks across current frontier models.
Governance
Map which user-side security protocols current models accept or refuse, and what that implies for public-sector procurement.
If you are a researcher or engineer with a concrete question in alignment, application-layer privacy, or governance, write to research@vitrlabs.no with a short note, a CV, and the dates you have in mind. We cannot host everyone; we reply to mail that is specific.