I am currently a Data Scientist at the Wharton School of the University of Pennsylvania.  

 

I work with research teams at Wharton in researching and developing data-driven solutions to complex and high-stake real-world problems, including detecting biases in human and algorithmic decision-making, improving the interpretability of complex models, and adapting reinforcement learning to problems in new domains.

​

I have been focusing on improving and expanding social scientists' toolkits to conduct text analyses. My collaborators and I propose Transformer Encoder for Social Science (TESS), a pretrained deep neural network encoder that shows superior performance on social science text processing tasks.

​

I received my Ph.D. in Politics from Princeton University. My dissertation puts forward an alternative theoretical framework emphasizing how low asset mobility helps foreign firms gain government support. Before coming to Princeton, I obtained an M.A. from Duke University and a B.A. from Fudan University, China. 

​

Email: hge [AT] wharton.upenn.edu