Karthik Narasimhan
Computer scientist
Princeton University
Karthik Narasimhan is a computer scientist at Princeton University and the senior academic on the original GEO paper. He coauthored OpenAI's first GPT paper and later work on language agents such as ReAct and Tree of Thoughts, so his work sits both at the naming of generative engine optimization and upstream of agentic search.
From OpenAI to Princeton
Narasimhan's research career spans the modern language-model stack. He coauthored OpenAI's early GPT work, then built an academic lab at Princeton focused on natural language processing, reinforcement learning, and language agents that can reason and use tools.
Senior Author on the GEO Paper
As senior academic on GEO: Generative Engine Optimization with lead author Pranjal Aggarwal and coauthors, Narasimhan helped frame visibility inside generative engines as a research problem with benchmarks and metrics. GEO-bench and the paper's measured content interventions gave practitioners an academic reference when commercial GEO claims were still mostly anecdotal.
ReAct and Tree of Thoughts
Before GEO became a marketing acronym, Narasimhan's group helped define how language agents combine reasoning with tools. ReAct showed models interleaving reasoning traces with actions such as search. Tree of Thoughts and related work extended deliberate multi-step reasoning. Those papers matter for GEO because agentic search is where models choose tools, fetch sources, and take steps on a user's behalf.
The systems marketers try to influence descend partly from research on agents and tool use that Narasimhan helped create.