PeopleofSuperintelligence

Researcher

Kirill Neklyudov

Also known as Kyrylo Nekliudov, Kirill O. Neklyudov

Kirill Neklyudov is a machine learning researcher who applies generative and Monte Carlo methods to problems in the natural sciences. Born in Sevastopol, Ukraine, he studied data analysis at Yandex and completed a 2020 dissertation at HSE on Bayesian deep learning. His work moved from Bayesian pruning and variance layers through MCMC and Action Matching to Wasserstein methods for quantum systems and population dynamics. Postdoctoral work took him to the University of Amsterdam and the Vector Institute. At Université de Montréal, he is an assistant professor and Core Academic Member at Mila, and he works at Anthropic while on leave.

Badges

Timeline

  1. By 2020

    Yandex and HSE training

    Born in Sevastopol, Ukraine, he starts competing in Ukrainian mathematics and physics olympiads at 13. He interns at Yandex from November 2013 to June 2014 and attends Yandex School of Data Analysis from 2014 to 2016. At HSE, he is a postgraduate student in the Faculty of Computer Science and Big Data and Information Retrieval School, and completes a 2020 dissertation on a Bayesian approach to improving discriminative and generative deep-learning models.

  2. 2017

    Structured Bayesian pruning

    In 2017, he coauthors Structured Bayesian Pruning via Log-Normal Multiplicative Noise, a Bayesian model that creates structured sparsity by removing neurons or convolutional channels from neural networks.

  3. By 2019

    Variance layers

    By 2019, he coauthors Variance Networks, which introduces stochastic layers whose weights follow zero-mean distributions parameterized by variance. Its layers support exploration in reinforcement-learning tasks and offer defense against adversarial attacks.

  4. 2020–2022

    Two MCMC frameworks

    In 2020, he coauthors Involutive MCMC, a framework that unifies many existing MCMC algorithms and suggests extensions. In 2022, he coauthors Orbital MCMC, which derives acceptance tests for arbitrary deterministic proposal maps and offers two algorithms based on periodic orbits and optimization trajectories.

  5. 2023

    Action Matching from samples

    In 2023, he coauthors Action Matching, which learns continuous system dynamics from independent samples over time. Its applications include experiments in biology, physics, and generative modeling.

  6. 2023–2024

    Quantum Monte Carlo and flow methods

    In 2023, he coauthors Wasserstein Quantum Monte Carlo, which uses a Wasserstein gradient flow and reaches molecular-system ground states faster in experiments. In 2024, he coauthors a deep-learning framework for Wasserstein Lagrangian flows that outperforms previous methods on single-cell trajectory inference.

  7. 2025–2026

    Diffusion and population methods

    In 2025, he coauthors a diffusion-model sampler for optimization within learned data manifolds when constraints are unknown. In 2026, he coauthors an algorithm that learns population mechanics from observed marginals and forecasts and interpolates unseen marginals.

  8. As of October 2026

    Assistant professor on leave

    His postdoctoral work takes him to the University of Amsterdam with Max Welling and the Vector Institute with Alán Aspuru-Guzik and Alireza Makhzani. He develops generative modelling, Monte Carlo methods and optimal transport for problems in natural sciences. As of October 2026, he is an assistant professor at the University of Montreal and a Core Academic Member at Mila. On leave, he works at Anthropic with Jascha Sohl-Dickstein.

Career

  1. NowUniversité de MontréalAssistant Professor
  2. NowMilaCore Academic Member
  3. NowAnthropic
  4. 2013–2014YandexIntern
  5. University of AmsterdamPostdoctoral researcher
  6. Vector InstitutePostdoctoral researcher

Education

  1. 2020National Research University Higher School of EconomicsPostgraduate student
  2. 2016Yandex

From the biography.

  1. The median h-index is 48 at universities and 14 at frontier AI labsIncludes Kirill Neklyudov. 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. 16% of AI career moves in the 2020s went from a university to a companyIncludes Kirill Neklyudov. We estimate that 16% of AI career moves in the 2020s went from a university to a company; 29% of moves starting at a university went to a company, and 76% ended the university role.
  3. 22 of 133 people shaping AI with a former OpenAI role now work at AnthropicIncludes Anthropic. Our estimates of where OpenAI recruits came from and where 133 people with former OpenAI roles work now, including the companies and labs they founded.
  4. Companies supply the most prior-work hires at 7 of the 9 frontier AI labsIncludes Anthropic. Our estimates of the organizations where 550 people worked directly before joining 9 frontier AI labs across 620 career transitions, comparing the main feeder organizations per lab.

Colleagues

People named in this biography, and people with dated positions at the same organization in the same years.

Sources and credits

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

  1. Kirill Neklyudov / (Kyrylo Nekliudov) necludov.github.io. Checked 21 October 2026.
  2. Staff - Kirill O. Neklyudov – HSE University hse.ru. Checked 21 October 2026.
  3. [1705.07283] Structured Bayesian Pruning via Log-Normal Multiplicative Noise arxiv.org. Checked 21 October 2026.
  4. [1803.03764] Variance Networks: When Expectation Does Not Meet Your Expectations arxiv.org. Checked 21 October 2026.
  5. Involutive MCMC: a Unifying Framework proceedings.mlr.press. Checked 21 October 2026.
  6. Orbital MCMC proceedings.mlr.press. Checked 21 October 2026.
  7. Action Matching: Learning Stochastic Dynamics from Samples proceedings.mlr.press. Checked 21 October 2026.
  8. Wasserstein Quantum Monte Carlo: A Novel Approach for Solving the Quantum Many-Body Schrödinger Equation proceedings.neurips.cc. Checked 21 October 2026.
  9. A Computational Framework for Solving Wasserstein Lagrangian Flows proceedings.mlr.press. Checked 21 October 2026.
  10. Diffusion Models as Constrained Samplers for Optimization with Unknown Constraints proceedings.mlr.press. Checked 21 October 2026.
  11. A Call to Lagrangian Action: Learning Population Mechanics from Temporal Snapshots proceedings.mlr.press. Checked 21 October 2026.