PeopleofSuperintelligence

Researcher

Sergey Shirobokov

Sergey Shirobokov is a Senior Research Engineer at Google DeepMind whose work spans machine learning for physics, algorithm discovery, formal proof search and computer architecture. He graduated in physics from Moscow State University in 2015, then pursued doctoral study at Imperial College London while studying data analysis at the Yandex School and working on SHiP's neutrino detector. He researched graph link prediction at Twitter and ShareChat, then helped develop AlphaEvolve in 2025. In 2026, he coauthored studies on formal proof search and AI-driven computer-architecture discovery.

Works
26
Citations
477
h-index
12
Badges

Publication numbers from OpenAlex, as of 22 October 2026.

Timeline

  1. 2015–2019

    Physics and SHiP research

    He graduates from the Department of Physics at Moscow State University in 2015. By 2019, he is a PhD student at Imperial College London and a student at the Yandex School of Data Analysis. At SHiP, he works on neutrino-detector simulation and applies machine learning to dark-matter searches.

  2. 2020

    Local generative surrogates

    In 2020, he coauthors Black-Box Optimization with Local Generative Surrogates, which uses deep generative models to approximate black-box simulators and enable gradient-based optimization.

  3. July 2021–January 2023

    Machine learning at Twitter

    He works as a machine-learning researcher in Twitter's Cortex Applied Research group from July 2021 to January 2023. He develops a scalable graph-neural-network algorithm for link prediction.

  4. 2022–2023

    Link prediction at ShareChat

    In February 2023, he joins ShareChat as a senior machine-learning researcher focused on large-scale applications for content discovery and engagement. His 2022 paper, Graph Neural Networks for Link Prediction with Subgraph Sketching, presents ELPH and BUDDY for scalable link prediction and appears at ICLR in 2023.

  5. 2025

    Helping develop AlphaEvolve

    He coauthors AlphaEvolve in 2025, a coding agent that uses large language models and evolutionary feedback to improve algorithms through direct code changes.

  6. 2026

    AI-assisted formal proof search

    In 2026, he coauthors a study of AI-driven formal proof search. Its most capable agent resolves 9 of 353 open Erdős problems and proves 44 of 492 OEIS conjectures.

  7. 2026

    AI for computer architecture

    He coauthors ArchAgent in February 2026, an automated architecture-discovery system built on AlphaEvolve that creates cache-replacement policies. Its first paper reports a policy with 5.3% IPC speedup over the prior state of the art on public multi-core Google traces. An August 2026 follow-up, ArchAgent v2, designs a three-level data prefetcher that beats the winning hand-designed solution at DPC4.

  8. Since February 2024

    Senior Research Engineer

    He joins Google DeepMind as a Senior Research Engineer in February 2024. As of October 2026, he is based in London.

Career

  1. Since 2024Google DeepMindSenior Research Engineer
  2. From 2023ShareChatSenior machine-learning researcher
  3. 2021–2023TwitterMachine-learning researcher

Education

  1. 2015Lomonosov Moscow State UniversityPhysics degree
  2. Imperial College LondonPhD

From the biography and OpenAlex affiliations (the years of papers that list the organization), so some years are approximate.

Most-cited works

  1. Sensitivity of the SHiP experiment to Heavy Neutral Leptons2019 · Journal of High Energy Physics · 114 citations
  2. Fast simulation of muons produced at the SHiP experiment using Generative Adversarial Networks2019 · Journal of Instrumentation · 54 citations
  3. The experimental facility for the Search for Hidden Particles at the CERN SPS2019 · Journal of Instrumentation · 49 citations
  4. The SHiP experiment at the proposed CERN SPS Beam Dump Facility2022 · The European Physical Journal C · 43 citations
  5. Sensitivity of the SHiP experiment to dark photons decaying to a pair of charged particles2021 · The European Physical Journal C · 38 citations

From OpenAlex, as of 21 October 2026.

  1. The median h-index is 48 at universities and 14 at frontier AI labsIncludes Sergey Shirobokov. The median h-index is 48 at universities and 14 at frontier AI labs. We compare publishing careers, recent citations, dual appointments and job moves across 2,420 university researchers and 308 frontier-lab researchers.
  2. 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 each year from 1955 to 2025, tracing the shift from university labs to companies.
  3. Google has the highest combined h-index among AI organizationsIncludes Google. We estimate that Google holds the highest combined h-index at 8,016, while universities account for 73% of the deduplicated total across AI organizations.
  4. 20% of people at companies who shape AI also hold a university role, led by Google and Stanford UniversityIncludes Google. We estimate that 539 of the 6,766 people with a known current organization (8%) hold both a university role and a company role, including 20% of people at companies and 13% of people at universities.

Colleagues

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

Sources and credits

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

  1. Mega Science fall 2019 research.misis.ru. Checked 20 October 2026.
  2. [2002.04632] Black-Box Optimization with Local Generative Surrogates arxiv.org. Checked 20 October 2026.
  3. README.md at main · SergeyShirobokov/Sergey-Shirobokov · GitHub github.com. Checked 20 October 2026.
  4. [2209.15486] Graph Neural Networks for Link Prediction with Subgraph Sketching arxiv.org. Checked 20 October 2026.
  5. AlphaEvolve: A Gemini-powered coding agent for designing advanced algorithms — Google DeepMind deepmind.google. Checked 20 October 2026.
  6. [2506.13131] AlphaEvolve: A coding agent for scientific and algorithmic discovery arxiv.org. Checked 20 October 2026.
  7. [2605.22763] Advancing Mathematics Research with AI-Driven Formal Proof Search arxiv.org. Checked 20 October 2026.
  8. [2602.22425] ArchAgent: Agentic AI-driven Computer Architecture Discovery arxiv.org. Checked 20 October 2026.
  9. [2608.09874] ArchAgent v2: A Case Study with the Data Prefetching Championship arxiv.org. Checked 20 October 2026.