PeopleofSuperintelligence

Researcher

Esmeralda S. Whitammer

Also known as Nikolay Malkin

Esmeralda Whitammer is a machine-learning researcher at the University of Edinburgh whose work on generative flow networks (GFlowNets) and amortized inference trains models to sample from unnormalized target distributions. A mathematician by training, she earned a Ph.D. in mathematics at Yale University in 2021 and then worked with Yoshua Bengio as a postdoctoral researcher at Mila in Montréal until 2024. Her papers include the trajectory balance objective for GFlowNets at NeurIPS 2022 and a theory of continuous GFlowNets at ICML 2023, and her ICLR 2024 paper on amortizing intractable inference in large language models earned an Honorable Mention. She became a Chancellor's Fellow in Edinburgh's School of Informatics in July 2024. Since April 2025, she is also a Programme Fellow at CIFAR in the Learning in Machines and Brains programme.

Timeline

  1. 2015–2021

    Mathematics at Washington and Yale

    Trained as a pure mathematician, she earns a B.S. at the University of Washington in 2015 and an M.S. and Ph.D. in mathematics at Yale University in 2021. Her Yale dissertation, advised by Alexander Goncharov in the Department of Mathematics, studies the Galois action on motivic fundamental groups of punctured elliptic and rational curves.

  2. 2021–2024

    Postdoc with Bengio at Mila

    From 2021 to 2024, she is a postdoctoral researcher at Mila – Québec AI Institute in Montréal and at the Université de Montréal's Department of Informatics and Operations Research. She works with Yoshua Bengio and collaborates with Aaron Courville, Gauthier Gidel and Guillaume Lajoie.

  3. 2022

    Trajectory balance for GFlowNets

    In 2022, she coauthors "Trajectory balance: Improved credit assignment in GFlowNets" with Moksh Jain, Emmanuel Bengio, Chen Sun and Yoshua Bengio, published at NeurIPS 2022. The paper proposes trajectory balance, a learning objective for generative flow networks that is more efficient than the earlier flow matching and detailed balance objectives.

  4. 2023

    A theory of continuous GFlowNets

    In 2023, she coauthors "A theory of continuous generative flow networks" at the 40th International Conference on Machine Learning, with Salem Lahlou, Tristan Deleu, Yoshua Bengio and others. GFlowNets are amortized variational inference algorithms that sample from unnormalized target distributions, and the paper presents a theory that extends them to continuous or hybrid state spaces.

  5. 2024

    ICLR 2024 honorable mention

    In 2024, the paper "Amortizing intractable inference in large language models", written with Edward J. Hu, Moksh Jain, Eric Elmoznino, Younesse Kaddar, Guillaume Lajoie and Yoshua Bengio, receives an Honorable Mention at ICLR 2024. It fine-tunes language models with GFlowNets so that they sample from intractable posterior distributions, and it treats chain-of-thought reasoning as a latent variable modeling problem.

  6. 2024

    Towards an AI mathematician

    In 2024, she and Yoshua Bengio publish the essay "Machine learning and information theory concepts towards an AI Mathematician" in the Bulletin of the American Mathematical Society. It takes an information-theoretic view of what makes a mathematical statement interesting, as a guide for future work on an AI mathematician.

  7. July 2024

    Chancellor's Fellow at Edinburgh

    In July 2024, she joins the University of Edinburgh's School of Informatics as a Chancellor's Fellow in AI and Data Science. She is a member of the school's Institute for Machine Learning.

  8. Since April 2025

    CIFAR Programme Fellow

    Since April 2025, she is a Programme Fellow at CIFAR in the Learning in Machines and Brains programme, and she remains a Chancellor's Fellow at Edinburgh. In spring 2026, she teaches Advanced Topics in Machine Learning, a course designed with Slava Borovitskiy and Rik Sarkar.

Career

  1. Since 2025CIFARProgramme Fellow
  2. Since 2024University of EdinburghChancellor's Fellow in AI and Data Science
  3. 2021–2024MilaPostdoctoral Researcher
  4. 2021–2024Université de MontréalPostdoctoral Researcher

Education

  1. 2021Yale UniversityM.S. in mathematics
  2. 2015University of WashingtonB.S.

From the biography.

  1. The Massachusetts Institute of Technology trained or employed 44 of the 187 AI major-prize winners, more than any other universityIncludes University of Edinburgh. 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.
  2. Universities had a net outflow of 2,068 moves in AI careersIncludes University of Edinburgh. Our estimates of career flows and median h-index across universities and companies in AI, including the role of joint appointments and earlier academic jobs.
  3. People shaping AI work at a median of 3 organizations over a career: 3 for university starters and institute and lab starters, versus 2 for company startersIncludes University of Edinburgh. Across 6,853 people shaping AI with a dated work role, we estimate a median of 3 parent organizations over a career (mean 3.6), comparing role groups, starting sectors and starting decades from the 1950s to 2020–2026.
  4. The company share of AI researchers’ top papers was 31% in 2020–2024 and 8% in 1970–1974Includes University of Edinburgh. We estimate where the people shaping AI worked when their most-cited papers appeared: the company share was 31% in 2020–2024 and 8% in 1970–1974, and Google led all organizations in 2020–2024.

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 Esmeralda S. Whitammer yet. Send us a free-licensed photo.

  1. Esmeralda S. Whitammer – personal website whitammer.github.io. Checked 22 October 2026.
  2. On the Galois Action on Motivic Fundamental Groups of Punctured Elliptic and Rational Curves – Yale Graduate School of Arts and Sciences Dissertations elischolar.library.yale.edu. Checked 22 October 2026.
  3. Esmeralda S. Whitammer – CIFAR cifar.ca. Checked 22 October 2026.
  4. Trajectory balance: Improved credit assignment in GFlowNets – arXiv arxiv.org. Checked 22 October 2026.
  5. A theory of continuous generative flow networks – Proceedings of Machine Learning Research proceedings.mlr.press. Checked 22 October 2026.
  6. ICLR 2024 Outstanding Paper Awards – ICLR Blog blog.iclr.cc. Checked 22 October 2026.
  7. Amortizing intractable inference in large language models – arXiv arxiv.org. Checked 22 October 2026.
  8. Machine learning and information theory concepts towards an AI Mathematician – arXiv arxiv.org. Checked 22 October 2026.
  9. Esmeralda Whitammer – School of Informatics, University of Edinburgh people.inf.ed.ac.uk. Checked 22 October 2026.