PeopleofSuperintelligence

Researcher

Pavlos Protopapas

Pavlos Protopapas is a Harvard researcher who applies machine learning to astronomy and scientific modeling. He studied physics at Imperial College in London, earned a theoretical physics PhD at the University of Pennsylvania, and served as associate director of its National Scalable Cluster Project. He developed a Harvard research and education collaboration with Chilean researchers in 2013 and coauthored NeuroDiffEq and astroBERT. By 2025, his work included transformer architectures for astronomical data. As of October 2026, he is Scientific Program Director and Lecturer at Harvard's Institute for Applied Computational Science.

Works
296
Citations
6,435
h-index
37

Publication numbers from OpenAlex, as of 27 October 2026.

Timeline

  1. 2005–2013

    Research at Harvard-Smithsonian

    From 2005 to 2013, he works at the Harvard-Smithsonian Center for Astrophysics. Across his research career, he applies machine learning, deep learning, and statistics to astronomy, with work in classification, probabilistic graphical modeling, and time-series analysis.

  2. 2013–2018

    Supernova research in Chile

    He develops a Harvard research and education collaboration with Chilean researchers in 2013. By 2018, students in its fifth year use machine learning to classify and visualize supernovae.

  3. By 2018

    Physics and grid computing

    By 2018, he serves as Scientific Program Director and Lecturer at Harvard's Institute for Applied Computational Science. A native of Cyprus, he joins the military after high school, studies physics at Imperial College in London, earns a theoretical physics PhD at the University of Pennsylvania, and serves as associate director of its National Scalable Cluster Project, an early grid-computing effort.

  4. 2020

    NeuroDiffEq for differential equations

    In 2020, he coauthors NeuroDiffEq, a Python package for solving differential equations with neural networks.

  5. 2021

    Building astroBERT for astronomy

    In 2021, he coauthors astroBERT, a language model trained with machine learning and natural language processing on recent astronomy publications. The project aims to improve NASA ADS discoverability and develop a named-entity recognition tool.

  6. By 2025

    Transformers for astronomical data

    By 2025, his research includes Astromer and Spectromer, transformer architectures for temporal and spectral modeling of astronomical data, plus machine-learning pipelines for the Event Horizon Telescope aimed at inferring black-hole spin and accretion dynamics.

  7. As of October 2026

    Scientific Program Director at Harvard

    As of October 2026, he is Scientific Program Director and Lecturer at Harvard's Institute for Applied Computational Science. He teaches Advanced Practical Data Science, a course on machine-learning operations and large-language-model applications.

Career

  1. NowHarvard UniversityScientific Program Director and Lecturer at the Institute for Applied Computational Science
  2. 2005–2013Center for Astrophysics Harvard & Smithsonian
  3. University of PennsylvaniaAssociate Director of the National Scalable Cluster Project

Education

  1. Imperial College London
  2. University of PennsylvaniaPhD

From the biography.

Checked career sources: samford.edu, coursebrowser.dce.harvard.edu.

Sources and credits

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

  1. Dr. Pavlos Protopapas adsabs.github.io. Checked 26 October 2026.
  2. Computational problems of cosmic proportions seas.harvard.edu. Checked 26 October 2026.
  3. Samford Professional Network Hosts Harvard University Professor and World-Renowned Data Scientist Pavlos Protopapas samford.edu. Checked 26 October 2026.
  4. Journal of Open Source Software: NeuroDiffEq: A Python package for solving differential equations with neural networks joss.theoj.org. Checked 26 October 2026.
  5. Building astroBERT, a language model for Astronomy & Astrophysics arxiv.org. Checked 26 October 2026.
  6. Pavlos Protopapas hdsr.mitpress.mit.edu. Checked 26 October 2026.
  7. Advanced Practical Data Science - Harvard Division of Continuing Education Course Browser coursebrowser.dce.harvard.edu. Checked 26 October 2026.