PeopleofSuperintelligence

Researcher

Łukasz Kaiser

Also known as Lukasz Kaiser

Łukasz Kaiser is a researcher whose papers span learned algorithms, language models, and software for experiments. With Ilya Sutskever, he studied neural networks that learned arithmetic procedures and were tested on longer inputs. He co-authored the Transformer and contributed to TensorFlow, Google’s neural translation system, and Tensor2Tensor. His first-author paper One Model To Learn Them All examined a network shared across language, speech, and image tasks. Universal Transformers explored repeated attention-based processing. His later shared work includes a 2025 paper on competitive programming with reasoning models. Together these projects study learned computation and ways to make it usable in research.

Now at
Google
Works
116

Publication numbers from OpenAlex, as of 14 October 2026. Citations and h-index are hidden because this OpenAlex record may be incomplete or mix in other authors.

Photo of Łukasz Kaiser

Timeline

  1. 2015

    Learning algorithms

    Kaiser and Sutskever published Neural GPUs Learn Algorithms in 2015.

  2. June 2017

    The Transformer

    Kaiser co-authored the Transformer paper in 2017.

  3. June 2017

    Learning across tasks

    Kaiser led the author list for One Model To Learn Them All in 2017.

  4. March 2018

    Tensor2Tensor

    Kaiser co-authored Tensor2Tensor in 2018.

  5. July 2018

    Repeated attention

    Kaiser co-authored Universal Transformers in 2018.

  6. February 2025

    Reasoning on programming problems

    Kaiser co-authored the 2025 competitive programming paper.

Sources and credits

Photo: Łukasz Kaiser GitHub profile. License: Not stated by source. Source: github.com. Cropped to a square and resized. Ask us to remove or replace this photo.

  1. Neural GPUs Learn Algorithms arxiv.org. Checked 11 October 2026.
  2. TensorFlow arxiv.org. Checked 11 October 2026.
  3. Google Neural Machine Translation arxiv.org. Checked 11 October 2026.
  4. Attention Is All You Need arxiv.org. Checked 11 October 2026.
  5. One Model To Learn Them All arxiv.org. Checked 11 October 2026.
  6. Tensor2Tensor for Neural Machine Translation arxiv.org. Checked 11 October 2026.
  7. Universal Transformers arxiv.org. Checked 11 October 2026.
  8. Competitive Programming with Large Reasoning Models arxiv.org. Checked 11 October 2026.