PeopleofSuperintelligence

Researcher

Karthik Narasimhan

Also known as Karthik R. Narasimhan

Karthik Narasimhan is a computer scientist whose work on language models and AI agents includes GPT, ReAct and SWE-bench. After a 2017 PhD at MIT, he spent a year at OpenAI contributing to the GPT language model and joined Princeton in 2018. A paper he coauthored won a NeurIPS 2022 Outstanding Paper award, and the National Science Foundation gave him a CAREER award in 2023. From 2023 to 2025 he was head of research at Sierra, where he introduced the τ-bench agent benchmark in 2024. He is an associate professor of computer science at Princeton, associate director of Princeton Language and Intelligence, and a co-founder of Curium.

Now at
Princeton University and Curium
Badges

Timeline

  1. 2015–2017

    A PhD at MIT

    He receives best paper recognition at EMNLP in 2015 and 2016. In 2017, he completes a PhD at MIT, where he works with Regina Barzilay at CSAIL on models that integrate reinforcement learning with language understanding.

  2. 2017–2018

    A year at OpenAI

    He spends a year at OpenAI as a visiting research scientist from 2017 to 2018, contributing to the GPT language model. The GPT paper, of which he is a coauthor, demonstrates an autoregressive transformer for language modeling and introduces the idea of solving NLP tasks through token prediction.

  3. 2018

    Joining Princeton

    He joins the Princeton computer science faculty in 2018, after his year at OpenAI. Princeton Engineering also gives him a commendation for outstanding teaching in 2018.

  4. 2022–2023

    Wentz and CAREER awards

    In 2022, Princeton Engineering gives him the Howard B. Wentz, Jr. Junior Faculty Award, which provides $50,000 to support research. In 2023, he receives a National Science Foundation CAREER award for the project "Reading to learn: Language-guided machine learning".

  5. 2022

    Outstanding paper at NeurIPS

    He is among the authors, with Sreejan Kumar, Thomas L. Griffiths and others, of "Using natural language and program abstractions to instill human inductive biases in machines", which wins an Outstanding Paper award at NeurIPS 2022. The paper shows that co-training meta-learning agents to predict representations of natural-language task descriptions and programs guides them toward more human-like inductive biases.

  6. 2022–2023

    ReAct and SWE-bench

    In 2022, he coauthors ReAct with Shunyu Yao and others, a method in which language models generate reasoning traces and task-specific actions in an interleaved manner. In 2023, he coauthors SWE-bench, an evaluation framework of 2,294 software engineering problems drawn from real GitHub issues across 12 popular Python repositories.

  7. 2023–2025

    Head of research at Sierra

    He is head of research at Sierra from 2023 to 2025. In June 2024, he introduces τ-bench with Sierra's research team, a benchmark for evaluating AI agents' performance and reliability in real-world settings with dynamic user and tool interaction.

  8. As of October 2026

    Professor and Curium co-founder

    He is an associate professor of computer science at Princeton, where he co-directs Princeton NLP and is associate director of Princeton Language and Intelligence, which Princeton launches in 2023. He is also a co-founder of Curium, a company whose founding team invented Generative Engine Optimization (GEO) for AI search.

Career

  1. Since 2018Princeton UniversityFaculty
  2. NowCuriumCo-founder
  3. 2023–2025SierraHead of Research
  4. 2017–2018OpenAIVisiting Research Scientist

Education

  1. 2017Massachusetts Institute of TechnologyPhD

From the biography.

  1. The Massachusetts Institute of Technology trained or employed 44 of the 187 AI major-prize winners, more than any other universityIncludes Karthik Narasimhan. The Massachusetts Institute of Technology trained or employed 44 of the 187 AI major-prize winners, more than any other university. We compare doctorate and 5+ year employment links, career timing and current workplaces for 187 major-prize and 1,111 best-paper winners.
  2. 11% of AI doctorate holders from 2010 to 2016 founded or led a company or lab by 10 years after graduationIncludes Karthik Narasimhan. We estimate how often AI doctorate holders founded or led a company or lab by 10 years after graduation, how waiting times compare between founders and top jobs, and which universities trained the most.
  3. 84% of people shaping AI with recorded undergraduate and doctoral schools changed institutions for their doctorateIncludes Princeton University. 84% of people shaping AI with recorded undergraduate and doctoral schools changed institutions for their doctorate. Across 2,116 people with both an undergraduate degree and a doctorate recorded, we trace which undergraduate schools feed which doctoral programs, how often students stay at the same school, and how paths differ by birth-country region.
  4. More people shaping AI studied at Massachusetts Institute of Technology than at any other institutionIncludes Princeton University. We estimate where the people shaping AI studied and where their careers led. Massachusetts Institute of Technology leads with 467 former students, including 291 recorded doctorates.

Colleagues

People named in this biography, and people with dated positions at the same organization in the same years.

Sources and credits

Photo: We have no published photo of Karthik Narasimhan yet. Send us a free-licensed photo.

  1. Karthik Narasimhan - Princeton Computer Science cs.princeton.edu. Checked 21 October 2026.
  2. Grounding natural language with autonomous interaction | CS cs.princeton.edu. Checked 21 October 2026.
  3. Karthik R. Narasimhan karthiknarasimhan.com. Checked 21 October 2026.
  4. Karthik Narasimhan | Sierra sierra.ai. Checked 21 October 2026.
  5. Danqi Chen and Karthik Narasimhan recipients of SEAS's Junior Faculty Award cs.princeton.edu. Checked 21 October 2026.
  6. Seven engineering faculty members receive NSF CAREER awards - Princeton Engineering engineering.princeton.edu. Checked 21 October 2026.
  7. NeurIPS 2022 Outstanding Paper Awards neurips.cc. Checked 21 October 2026.
  8. Using Natural Language and Program Abstractions to Instill Human Inductive Biases in Machines (arXiv:2205.11558) arxiv.org. Checked 21 October 2026.
  9. ReAct: Synergizing Reasoning and Acting in Language Models (arXiv:2210.03629) arxiv.org. Checked 21 October 2026.
  10. SWE-bench: Can Language Models Resolve Real-World GitHub Issues? (arXiv:2310.06770) arxiv.org. Checked 21 October 2026.
  11. τ-Bench: Benchmarking AI agents for the real-world | Sierra sierra.ai. Checked 21 October 2026.
  12. Beyond ChatGPT: Princeton Language and Intelligence initiative pushes the boundaries of large AI models | CS cs.princeton.edu. Checked 21 October 2026.
  13. Curium | Inventors of Generative Engine Optimization curium.ai. Checked 21 October 2026.