Dani Roytburg
e-mail: droytbur@andrew.cmu.edu or djroytburg@gmail.comMachine Learning Department, Carnegie Mellon University.
Research Fellow, Martian Research.
I am a Master's student in the Machine Learning Department at Carnegie Mellon University and Research Fellow at Martian Research. I am interested in building reliability mechanisms for trustworthy artificial intelligence. I am advised by Daphne Ippolito.
Currently, I study scalable oversight and mechanistic interpretability. I am actively working on multi-agent training paradigms for legible chain of thought, and on building activation probes for steering the behavior of LLM judges.
Previously at Emory University, I graduated summa cum laude with a Bachelor's in Computer Science and Statistics. I worked with Lauren Klein on graph-text modelling for digital humanities, as well as auditing biases in NLP techniques. I did my honor's thesis on structured prediction and argument mining under the supervision of Joyce Ho and Prof. Klein. I've also worked as a consultant with ZS Associates, Cloverpop, and McCormick, applying novel NLP-based analytics engines for commercial data science.