
RLHF & Human Values
Investigating the human values embedded in RLHF datasets
I am currently a Researcher (PhD Candidate) at Purdue University. Before coming to Purdue, I was a Research Engineer at the Red Lab, Indiana University. My research interests are broadly at the intersection of Neurosymbolic Reasoning, Robot Learning, Human Robot-Agent Interaction, and Safety Alignment. My long-term research goal is to build AI Agents and Robotic systems that can perceive, understand, and interact with the world by thinking about action-impact consequences. My research draws from multi-disciplinary fields to realize this goal. My work has been published in different computing venues, including NeurIPS D&B (Spotlight Paper), ICRA, IROS, CHI, CSCW, and ICSE. I have been fortunate to collaborate with both industry and government partners, including Microsoft and US EDA.
Investigating the human values embedded in RLHF datasets
Documenting the human values embedded into AI models.
Investigating the evolution of the #darkpatterns discourse on Twitter
A specialized data-intensive analytic platform developed for the US EDA.
Navigating the ethical complexities in engineering practice.
Investigating how open source AI enthusiasts appropriate AI models.