PeopleofSuperintelligence

Researcher

David Alvarez-Melis

David Alvarez-Melis is a computer scientist whose research connects optimal transport, data-centric machine learning, and interpretable AI. After mathematics degrees at ITAM and NYU, he worked in IBM Research's Speech Recognition Group and earned a computer science PhD at MIT. In 2018, he coauthored a paper on self-explaining neural networks and a 2021 paper on dataset transformation with Wasserstein gradient flows. Harvard SEAS appointed him an assistant professor in 2022, and in 2024 he joined the Kempner Institute as associate faculty. He is an assistant professor of computer science at Harvard SEAS, where he leads the Data-Centric Machine Learning group; he also works as a researcher at Microsoft Research New England.

Now at
Microsoft Research and Harvard John A. Paulson School of Engineering and Applied Sciences
Works
87
Badges

Publication numbers from OpenAlex, as of 27 October 2026. Citations and h-index are hidden because this OpenAlex record may be incomplete or mix in other authors.

Timeline

  1. 2006–2013

    Mathematics degrees at ITAM and NYU

    From 2006 to 2011, he studies applied mathematics at ITAM and earns a BSc. At Courant Institute, NYU, he completes an MS in mathematics in 2013.

  2. 2013–2014

    Speech recognition at IBM

    From 2013 to 2014, he works at IBM Research as a Supplemental Researcher in the Speech Recognition Group.

  3. 2014–2019

    A PhD on optimal transport

    At MIT CSAIL, he completes a PhD in computer science in 2019 under advisor Tommi Jaakkola. His thesis extends optimal transport to model structured data and invariances, with NLP applications including unsupervised word translation, sentence similarity, domain adaptation, and WordNet matching.

  4. 2018

    Self-explaining neural networks

    In 2018, he coauthors Towards Robust Interpretability with Self-Explaining Neural Networks at NeurIPS. The paper puts explanation into the learning process and sets three goals for explanations: explicitness, faithfulness, and stability.

  5. 2019–2021

    A postdoc at Microsoft Research

    In 2019, he joins Microsoft Research New England as a postdoctoral researcher.

  6. Since 2021

    Senior researcher at Microsoft

    He becomes a Senior Researcher at Microsoft Research New England in 2021, in its Machine Learning and Statistics group. His data-centric approach aligns and compares datasets.

  7. 2021

    Dataset transformation through optimal transport

    With Nicolò Fusi, he publishes Dataset Dynamics via Gradient Flows in Probability Space in 2021. The paper models dataset transformation with Wasserstein gradient flows and tests the method on classification constraints and transfer learning.

  8. 2022

    Joining Harvard SEAS

    In 2022, he joins Harvard SEAS as Assistant Professor of Computer Science and Applied Mathematics from Microsoft Research New England, where he is a Senior Researcher in the Machine Learning and Statistics group.

  9. July 2024

    Associate faculty at Kempner

    In July 2024, he joins the Kempner Institute as an associate faculty member while serving as a Harvard assistant professor of computer science. His research addresses learning in constrained, dynamic, and complex environments.

  10. Since 2023

    Assistant professor at Harvard

    In 2026, he is an assistant professor of computer science at Harvard SEAS and leads the Data-Centric Machine Learning group. He also works as a researcher at Microsoft Research New England.

Career

  1. Since 2022Harvard UniversityAssistant Professor of Computer Science and Applied Mathematics
  2. Since 2021Microsoft ResearchSenior Researcher
  3. NowHarvard John A. Paulson School of Engineering and Applied SciencesGroup Lead
  4. 2013–2014IBM ResearchSupplemental Researcher

Education

  1. 2019MIT CSAILPhD
  2. 2013Courant Institute of Mathematical SciencesMS in Mathematics
  3. 2011Autonomous Technological Institute of MexicoBSc in Applied Mathematics

From the biography.

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 David Alvarez-Melis yet. Send us a free-licensed photo.

  1. David Alvarez-Melis | Home dmelis.github.io. Checked 26 October 2026.
  2. David Alvarez-Melis at Microsoft Research microsoft.com. Checked 26 October 2026.
  3. Thesis Defense: David Alvarez-Melis: Optimal Transport in Structured Domains: Algorithms and Applications csail.mit.edu. Checked 26 October 2026.
  4. Towards Robust Interpretability with Self-Explaining Neural Networks papers.nips.cc. Checked 26 October 2026.
  5. Harvard Engineering Welcomes Ten New Faculty seas.harvard.edu. Checked 26 October 2026.
  6. Dataset Dynamics via Gradient Flows in Probability Space proceedings.mlr.press. Checked 26 October 2026.
  7. Kempner Institute appoints two new associate faculty members news.harvard.edu. Checked 26 October 2026.
  8. David Alvarez-Melis | Harvard John A. Paulson School of Engineering and Applied Sciences seas.harvard.edu. Checked 26 October 2026.