Hi!
My research answers how machine learning can be applied to explain and predict political behavior at scale. I focus on developing classification tools that combine experimentation, eye-tracking, and deep learning to understand how stable societal divisions influence political decisions. My work seeks to expand our understanding of these dynamics and improve predictive accuracy in political science, with a primary focus on the U.S. and post-Soviet space.
I prefer to think that my research mainly focuses on one or several of the following fields:
CPB
)PC
)AP
)
I plan/hope [insert degree of uncertainty here] to be on the job market in 2025-26.
AP
) (CPB
) (JMP-2)In a large-scale eye-tracking study, we show how motivated visual processing reinforces and potentially deepens partisan divides.
AP
) (PC
) (CPB
)We explore how media outlets from different ideological perspectives visually frame a polarizing issue, immigration, and how their target audiences are likely to interpret these frames. We illustrate yet again that media bias is important to study, but this time from the perspective of a visual slant.
We highlight a key limitation in studies that use large-scale approaches to tackle current challenges in visual sentiment analysis. In addition to demonstrating that certain biases cannot be fully addressed, we offer a guide on how to move forward.
CPB
) (PC
) (JMP-1)By using machine learning tools, I am exploring persistent structural properties of political propaganda canvas that (could) affect persuasion in autocracy. The project is in the data mining phase.
To explain why citizens support or oppose dictatorial leadership, most approaches to studying political behavior and public opinion in autocracies focus on mass explicit preferences. I show that implicit preferences - people's gut attitudes about politics - represent a significant and yet understudied source of citizens' disruptive preferences against the regime and in support of dictators' opponents.
The rest is in my CV.
Articles
PC
) (CPP
)CPP
)CPP
)Book chapters
CPP
)CPP
)