Published multiple papers at top-tier conferences (ICLR, NeurIPS, ICML) on topics including LLM planning, knowledge graph training, memorization, and transfer learning
Contributed to major projects such as Gemini 1.5
Background
Staff Research Scientist at Google DeepMind
Leads the DeepPhenomena team
Research focuses on enabling robust planning and scientific reasoning in Foundational Models
Systematically deconstructs reasoning failures in models to uncover underlying mechanisms and develop targeted interventions
Over a decade of research experience in the science of deep learning