PeopleofSuperintelligence

Researcher

Murat A. Erdogdu

Murat A. Erdogdu is a statistician and computer scientist whose theory work designs and analyzes optimization and sampling algorithms for machine learning models. He studied electrical engineering and mathematics at Bogazici University, earned a computer science MSc and a statistics PhD at Stanford, and spent 2017–2018 as a postdoctoral researcher at Microsoft Research New England. At the University of Toronto from 2018, he became a Canada CIFAR AI Chair at the Vector Institute, won a Connaught New Researcher Award in 2019, and visited the Simons Institute in 2021. Since 2025, he is an Associate Professor with tenure in computer science, and his recent papers study post-training with policy gradients and the training of neural networks.

Badges

Timeline

  1. 2011–2017

    Degrees at Bogazici and Stanford

    He earns bachelor's degrees in electrical engineering and mathematics at Bogazici University in 2011, an MSc in computer science at Stanford in 2015, and a PhD in statistics at Stanford in 2017. His dissertation, Stein's lemma and subsampling in large-scale optimization, uses tools from statistics to design scalable optimization algorithms, with Mohsen Bayati and Andrea Montanari as advisors.

  2. 2015

    Sub-sampled Newton methods

    In 2015, he and Andrea Montanari publish Convergence rates of sub-sampled Newton methods at NeurIPS. The paper combines sub-sampling with low-rank approximation in a randomized batch algorithm whose convergence rate is comparable to Newton's method at a much smaller per-iteration cost.

  3. 2017–2018

    Postdoc at Microsoft Research

    He is a postdoctoral researcher at Microsoft Research New England from 2017 to 2018.

  4. 2018

    Assistant professor at Toronto

    In 2018, he joins the University of Toronto as an assistant professor of computer science and of statistics, and becomes a faculty member of the Vector Institute for AI. CIFAR appoints him a Canada CIFAR AI Chair in 2018.

  5. 2018–2019

    Diffusions for optimization and sampling

    At NeurIPS 2018, he, Lester Mackey and Ohad Shamir show that the global-minimizer guarantee of the discretized Langevin diffusion holds for any suitably smooth diffusion. At NeurIPS 2019, a paper with Xuechen Li, Yi Wu and Lester Mackey establishes convergence rates for Langevin sampling discretized with stochastic Runge-Kutta, improving the best-known rate for strongly log-concave sampling with the overdamped Langevin equation.

  6. 2019

    A Connaught New Researcher Award

    He receives a Connaught New Researcher Award in 2019.

  7. 2021

    Visiting scientist at Simons Institute

    From August to December 2021, he is a visiting scientist at the Simons Institute in Berkeley, in the program Geometric Methods in Optimization and Sampling.

  8. Since 2025

    Associate Professor with tenure

    The Department of Computer Science promotes him to Associate Professor with tenure effective July 2025. His 2025 papers include Post-Training with Policy Gradients: Optimality and the Base Model Barrier and Learning Quadratic Neural Networks in High Dimensions: SGD Dynamics and Scaling Laws. He is also a faculty member of the Vector Institute, where his main interest is designing optimization algorithms for machine learning models.

Career

  1. Since 2018University of TorontoAssistant Professor
  2. Since 2018Vector InstituteFaculty Member
  3. 2021University of California, BerkeleyVisiting Scientist
  4. 2017–2018Microsoft ResearchPostdoctoral Researcher

Education

  1. 2017Stanford UniversityPhD in statistics
  2. 2015Stanford UniversityMSc in computer science
  3. 2011Boğaziçi UniversityBachelor's degrees in electrical engineering and mathematics

From the biography.

  1. 23% of people shaping AI who left a university for a company later returned to a universityIncludes Murat A. Erdogdu. We estimate that 468 of the 2,011 people shaping AI who left a university for a company (23%) later returned to a university after a median of 5 years, and 163 of 829 (20%) made the reverse round trip.
  2. The median h-index is 48 at universities and 14 at frontier AI labsIncludes University of Toronto. 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.
  3. Of 943 people shaping AI who worked at Google, 446 work there nowIncludes Vector Institute. Our estimates of where the 943 people shaping AI who worked at Google and Google DeepMind work now: 446 at Google and 452 across 278 other organizations.
  4. The Massachusetts Institute of Technology trained or employed 44 of the 187 AI major-prize winners, more than any other universityIncludes University of Toronto. The Massachusetts Institute of Technology trained or employed 44 of the 187 AI major-prize winners, more than any other university. We compare doctorate and 5+ year employment links, career timing and current workplaces for 187 major-prize and 1,111 best-paper winners.

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 Murat A. Erdogdu yet. Send us a free-licensed photo.

  1. Murat A. Erdogdu: academic employment, affiliations and education cs.toronto.edu. Checked 21 October 2026.
  2. Stein's lemma and subsampling in large-scale optimization | Stanford Digital Repository purl.stanford.edu. Checked 21 October 2026.
  3. Convergence rates of sub-sampled Newton methods papers.nips.cc. Checked 21 October 2026.
  4. Murat Erdogdu – CIFAR cifar.ca. Checked 21 October 2026.
  5. Global Non-convex Optimization with Discretized Diffusions papers.nips.cc. Checked 21 October 2026.
  6. Stochastic Runge-Kutta Accelerates Langevin Monte Carlo and Beyond papers.nips.cc. Checked 21 October 2026.
  7. Murat Erdogdu | Simons Institute for the Theory of Computing simons.berkeley.edu. Checked 21 October 2026.
  8. Department of Computer Science announces promotion of five faculty members web.cs.toronto.edu. Checked 21 October 2026.
  9. Murat A. Erdogdu cs.toronto.edu. Checked 21 October 2026.
  10. Murat A. Erdogdu - Vector Institute for Artificial Intelligence vectorinstitute.ai. Checked 21 October 2026.