PeopleofSuperintelligence

Researcher

Arvind Neelakantan

Arvind Neelakantan is a research scientist whose work runs from neural programs to language models. He co-wrote the Neural Programmer paper with Quoc Le and Ilya Sutskever, published at ICLR 2016. He co-authored the GPT-3 paper, 'Language Models are Few-Shot Learners', which NeurIPS named an Outstanding Paper in 2020. He now works on Gemini at Google DeepMind.

Timeline

  1. November 2015

    Gradient noise for deep networks

    In November 2015, he co-writes 'Adding Gradient Noise Improves Learning for Very Deep Networks' with Luke Vilnis, Quoc Le, Ilya Sutskever and three others.

  2. March 2016

    A Google PhD fellowship

    In March 2016, Google names him one of its PhD Fellows, from the University of Massachusetts, Amherst.

  3. 2016

    Neural Programmer at ICLR

    He co-writes 'Neural Programmer: Inducing Latent Programs with Gradient Descent' with Quoc Le and Ilya Sutskever, and the paper appears at the International Conference on Learning Representations in 2016.

  4. 2020

    GPT-3 and an outstanding paper

    He co-authors 'Language Models are Few-Shot Learners', the GPT-3 paper, and NeurIPS names it an Outstanding Paper in 2020.

  5. January 2022

    Text and code embeddings

    He co-writes 'Text and Code Embeddings by Contrastive Pre-Training', submitted to arXiv in January 2022.

  6. May 2023

    Lecturing on ChatGPT

    He gives a UCSB special lecture as an OpenAI speaker, 'Unraveling the Secrets of ChatGPT', in May 2023.

  7. Since 2025

    Working on Gemini

    He is a research scientist at Google DeepMind, where he works on Gemini.

Sources and credits

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

  1. [1511.06807] Adding Gradient Noise Improves Learning for Very Deep Networks arxiv.org. Checked 14 October 2026.
  2. Announcing the 2016 Google PhD Fellows for North America, Europe and the Middle East research.google. Checked 14 October 2026.
  3. Neural Programmer: Inducing Latent Programs with Gradient Descent research.google. Checked 14 October 2026.
  4. NeurIPS 2020 Awards neurips.cc. Checked 14 October 2026.
  5. [2005.14165] Language Models are Few-Shot Learners arxiv.org. Checked 14 October 2026.
  6. [2201.10005] Text and Code Embeddings by Contrastive Pre-Training arxiv.org. Checked 14 October 2026.
  7. Special Lecture | UCSB Center for Responsible Machine Learning ml.ucsb.edu. Checked 14 October 2026.
  8. Arvind Neelakantan | alphaXiv alphaxiv.org. Checked 14 October 2026.