I am an applied scientist focused on modeling design for large-scale decision and feedback systems, with an emphasis on measurable impact in search and advertising. My work involves identifying appropriate mathematical structures for ambiguous business problems, drawing on tools such as off-policy evaluation, temporal point processes, and contrastive estimation. I am also interested in understanding how multi-turn interaction, algorithmic distillation, and world-knowledge association contribute to emergent capabilities in language models. I frequently engage in early-stage evaluation of new ideas to assess feasibility and scientific soundness prior to execution.