PeopleofSuperintelligence

Researcher

Borislav Kozlovskii

Borislav Kozlovskii is a researcher whose work spans transformer adaptation, multilingual entity linking, quantum error correction and AI-assisted algorithm discovery. He coauthored a 2021 study on vocabulary transfer for transformers and a 2023 paper introducing BELA, a model for linking entities in 97 languages. In December 2024, he contributed to a Willow surface-code study; in 2025, he helped develop Google DeepMind's AlphaEvolve and coauthored research on unstable fluid singularities. By May 2026, AlphaEvolve reached applications in physics, electricity grids and computing infrastructure; in August 2026, he coauthored research that uses machine learning and AlphaEvolve to improve a matrix-multiplication bound.

Works
11
Citations
810
h-index
4

Publication numbers from OpenAlex, as of 19 October 2026.

Timeline

  1. 2021

    Transformer vocabulary transfer

    In 2021, he coauthors a study that tests whether corpus-specific tokenization improves transformer fine-tuning and reports faster transfer with stronger model performance.

  2. 2023

    BELA entity linking

    In 2023, he coauthors BELA, a model that detects and links entities in texts in 97 languages.

  3. December 2024

    Surface-code memories

    In December 2024, he contributes to a Willow study, which demonstrates distance-7 and distance-5 surface-code memories operating below threshold.

  4. May 2025

    Building AlphaEvolve

    In May 2025, he helps develop Google DeepMind's AlphaEvolve, a Gemini-powered coding agent that evolves algorithms for science and computing.

  5. September 2025

    AI and fluid singularities

    In September 2025, he coauthors research that uses AI methods to discover new families of unstable singularities in fluid equations.

  6. May 2026–August 2026

    AlphaEvolve and matrix multiplication

    By May 2026, AlphaEvolve, which he helped develop at Google DeepMind, has applications in physics, electricity grids and computing infrastructure. In August 2026, he coauthors work that combines machine learning with AlphaEvolve to improve a bound for the matrix-multiplication exponent.

Most-cited works

  1. Quantum error correction below the surface code threshold2024 · Nature · 762 citationsWith Edward Farhi and Michel Devoret
  2. Fine-tuning transformers: Vocabulary transfer2023 · Artificial Intelligence · 18 citations
  3. AlphaEvolve: A coding agent for scientific and algorithmic discovery2025 · 15 citationsWith Abbas Mehrabian, Adam Zsolt Wagner, Matej Balog, Pushmeet Kohli, Sebastian Nowozin and Swarat Chaudhuri
  4. Rotational energy surface and quasiclassical analysis of the rotational energy level cluster formation in the ground vibrational state of PH32007 · Journal of Molecular Spectroscopy · 9 citations
  5. Multilingual End to End Entity Linking2023 · 4 citations

From OpenAlex, as of 19 October 2026.

  1. The median h-index is 54 at universities and 19 at frontier AI labsIncludes Borislav Kozlovskii. The median h-index is 54 at universities and 19 at frontier AI labs. We compare publishing careers, citations and moves to estimate how universities and frontier labs contribute to AI research.
  2. Google has the highest combined h-index among AI organizationsIncludes Google. Our estimates of combined and median h-index at AI organizations, using current affiliations and OpenAlex records.
  3. Google alumni founded or led 81 other organizations, the most among the labs we compareIncludes Google. We trace 194 organizations founded or led by AI lab alumni, and 202 career moves between the labs.
  4. Google had the most people shaping AI every year from 2014 to 2025Includes Google. Our estimates of which organizations employed the most people shaping AI from 1961 to 2025, with a year-by-year chart.

Colleagues

People with dated positions at the same organization in the same years.

Sources and credits

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

  1. Fine-Tuning Transformers: Vocabulary Transfer arxiv.org. Checked 18 October 2026.
  2. Multilingual End to End Entity Linking arxiv.org. Checked 18 October 2026.
  3. Quantum error correction below the surface code threshold nature.com. Checked 18 October 2026.
  4. AlphaEvolve: A Gemini-powered coding agent for designing advanced algorithms deepmind.google. Checked 18 October 2026.
  5. AlphaEvolve: A coding agent for scientific and algorithmic discovery arxiv.org. Checked 18 October 2026.
  6. Discovering new solutions to century-old problems in fluid dynamics deepmind.google. Checked 18 October 2026.
  7. Discovery of Unstable Singularities arxiv.org. Checked 18 October 2026.
  8. AlphaEvolve: How our Gemini-powered coding agent is scaling impact across fields deepmind.google. Checked 18 October 2026.
  9. Improving the matrix multiplication exponent with modern optimization and AlphaEvolve arxiv.org. Checked 18 October 2026.