Research
Broadly, my research sits between computational social science and the analysis of authoritarian influence operations. I use large-scale text methods to characterise how (para-)state actors build and contest strategic narratives, and to make the structure of those campaigns legible at scale. I focus on Russia and the wider Russophone information space.
Russian strategic frames in Moldova’s 2024 elections
Undergraduate thesis in politics, University of Oxford (2026). Ranked 1st in cohort, awarded ‘Excellent’ First Class (82).
To my knowledge, this is the first cross-corpus computational frame analysis of Moldova’s 2024 electoral cycle: 321,241 messages from 53 mostly pro-Russian Telegram channels, paired with 15,176 alignment-coded articles from the Moldovan press. Pro-Kremlin messaging resolves into a four-frame discursive architecture: elite contestation dominant at 42.7 per cent of coded volume, ahead of the three imperial-identity discourses together (36.5 per cent), plus an off-theme residual (20.8 per cent) that is a descriptive finding in its own right. The scheme extends Tolz and Hutchings’s three-discourse account of Russophone oppositional media with elite contestation as a fourth substantive frame.
A documented, derived-form release of the corpus (channel list, message identifiers, topic and frame labels, rehydration code) is planned.
Do algorithms amplify identity?
M.A. research project (proposed), Department of Political Science, McGill University, 2026–. Supervisor: Aaron Erlich.
Pro-Kremlin content circulates across the former Soviet Union on the same platforms with markedly different political effects. This project audits one understudied mechanism behind that variation: algorithmic amplification. Monitored accounts calibrated to typical users in four countries track what recommendation systems push, how quickly, and to whom, asking whether platforms systematically favour identity-based content, and whether that varies across national contexts. It extends the Moldova pipeline from measuring frame supply to measuring frame distribution.