I’m a post-doctoral researcher at Stanford, working with Diyi Yang at the SALT Lab. I recently graduated with a PhD from the Center for Data Science at New York University where I was advised by Prof. He He.

I’m interested in measuring and mitigating the societal impacts of AI. This involves designing metrics that quantify effects such as homogenization in LLM outputs and user reliance on coding agents, evaluating these with both automatic evaluation and user studies, and then proposing targeted training methods to mitigate these. I’m currently thinking about (1) how to mitigate user overreliance on AI, (2) how AI agents shape the future of work and education, (3) how to support co-creativity with foundation models, (4) how we can elicit more diverse and novel LLM outputs for applications like scientific discovery.

If my name is familiar but you can’t place it, that might be from when I helped organize the NYU NLP and Text-as-Data talk series for many years. I’m also currently co-organizing the SALT & Talk series at Stanford. Prior to this, I completed my MS in Computer Science at NYU’s Courant Institute of Mathematical Sciences during which I was a Graduate Research Associate at the Center for Social Media and Politics working on applying deep learning to computational social science tasks. I did my undergrad at the National Institute of Technology - Karnataka where my thesis was advised by Prof. Sowmya Kamath.

Here’s my list of publications, CV and Google Scholar page. Also experimenting with a fun visualization of my work inspired by Phillipe’s research garden and Katarzyna’s research landscape. Head over to my personal blog for more light hearted content. All the other relevant links are in the footer. You can contact me at vishakkp@stanford.edu or vishakh@nyu.edu.

Updates:

  • September 2026: Offloading-Score was accepted as an Oral at NeurIPS, one of 15 papers selected in the Evaluations track.
  • July 2026: Two papers at COLM on routing for sample diversity and coding agent usage. See you in SF (well this need not be just at COLM :))!
  • July 2026 : Helping organize the [CIVIC-AI workshop](https://civic-ai-collaboration.github.io/) in Singapore.
  • May 2026: New paper, Offloading-Score, where we present a new way to quantify reliance on AI in coding agent workflows.
  • April 2026: New dataset, SWE-Chat, and paper to help us understand how real users work with coding agents.
  • April 2026: New pre-print on routing across models to improve sample diversity in open-ended generation.
  • February 2026: New pre-print with the SALT Lab on adaptive semi-structured interviewing for qualitative insight discovery.
  • January 2026: Our paper on LLM novelty will appear at ICLR 2026, see you in Brazil!
  • November 2025: Second pre-print in collaboration with the Midjourney Storytelling Lab, understanding the literary taste of users in a data-driven way.
  • September 2025: Moved to Stanford to join Diyi's SALT lab!
  • July 2025: My AI2 internship work on creating schemas to compare research papers was accepted to EMNLP Findings. Check out the paper here.
  • May 2025: My internship work from Adobe got accepted to ACL 2025. Check out the paper here . I think it has cool implications for how we should think and design agentic LLM pipelines.
  • April 2025: Passed my thesis defense, still can't quite believe it!
  • January 2025: Gave a talk covering a lot of my thesis work, titled 'Side Effects May Include Homogenization and Overusing Cliches' discussing how the way we do post training of LLMs today can have uninenteded side effects. The talk is available online here.
  • September 2024: Interning with the Semantic Scholar research team over the Fall.

  • June 2024: Attended Creativity and Cognition in Chicago, my first HCI conference, and presented our work on analyzing perspectives from emerging professional writers.

  • May 2024: I interned with the Document Intelligence Lab at Adobe in San Jose over the summer, working on long document generation prioritizing diversity of content with Jennifer Healey, David Arbour and Tong Sun.

  • May 2024: I attended ICLR 2024 to present our work on diversity in collaborative writing. On the way back, I also gave a talk about our work at Oxford in the UK!

  • March 2024: I gave virtual talks at Bocconi University and the University of Melbourne

  • December 2023: I attended EMNLP 2023 and helped give a tutorial on Creative Natural Language Generation with Tuhin Chakrabarty, Violet Peng and He He. See the slides here!

  • July 2023: I presented our work on extrapolative generation with iterative refinement at ICML 2023.

  • July 2023: I was the co-chair for the ACL 2023 Student Research Workshop along with Gisela Vallejo and Yao Fu.

    • </ul>