Karthik Narasimhan

• Program

  AI Safety Science

• Institution

  Princeton University

• Location

  USA

Dr. Karthik Narasimhan’s research focuses on problems at the intersection of language and decision making. He builds autonomous agents that learn to operate in the world using both their own experience and existing human knowledge. Karthik received his PhD from MIT in 2017, and spent a year as a visiting research scientist at OpenAI, contributing to the GPT language model, prior to joining Princeton in 2018. He is the recipient of a Google Research Scholar Award (2022), an Amazon research award (2019) and best paper awards/nominations at EMNLP (2015, 2016).
This project will create benchmarks, tests, and metrics to assess the reliability and consistency of AI software engineering (SWE) agents. It will specifically focus on identifying potential catastrophic failures, similar to the approach used by ToolEmu for evaluating LLM tool use.