PeopleofSuperintelligence

Researcher

Raja Marjieh

Raja Marjieh is a cognitive scientist who uses large language models and large-scale human experiments to study how people and machines represent similarity. He studied physics and electrical engineering at the Technion and earned a psychology PhD at Princeton, advised by Tom Griffiths. In 2022, he coauthored "Using natural language and program abstractions to instill human inductive biases in machines", which received a NeurIPS Outstanding Paper award, and in 2023 he was first author of an ICLR paper that tested 611 pre-trained models against human similarity judgments. His 2024 Scientific Reports paper showed that GPT models predict human sensory judgments across six psychophysical datasets. He is a Research Fellow at Harvard's Kempner Institute and is due to join the Tufts psychology department as an assistant professor in academic year 2027–28.

Timeline

  1. 2020

    Gibbs Sampling with People

    In 2020, he coauthors the NeurIPS oral paper "Gibbs Sampling with People" with Peter Harrison, Nori Jacoby and others. In the method, a participant uses a slider to move one stimulus dimension in each iteration, and the team tests it on musical chords, vocal emotions and faces.

  2. 2022

    A NeurIPS Outstanding Paper

    The paper "Using natural language and program abstractions to instill human inductive biases in machines" receives a NeurIPS 2022 Outstanding Paper award, and he is one of its ten authors, with Sreejan Kumar, Ishita Dasgupta, Tom Griffiths and others. Co-training meta-learning agents on predicting representations from natural language task descriptions and induced programs guides them toward more human-like inductive biases.

  3. 2023

    Language as a stand-in for similarity

    In 2023, he is first author of the ICLR paper "Words are all you need? Language as an approximation for human similarity judgments", written with Pol van Rijn, Ilia Sucholutsky, Theodore Sumers, Harin Lee, Tom Griffiths and Nori Jacoby. The study evaluates 611 pre-trained models on images, audio and video, finds a large gap from human similarity judgments, and proposes language-based approximations; the team releases 206,339 human judgments.

  4. February 2024

    Timbre and musical consonance

    In February 2024, Nature Communications publishes "Timbral effects on consonance disentangle psychoacoustic mechanisms and suggest perceptual origins for musical scales". He and Peter Harrison share first authorship, with Nori Jacoby and colleagues at the Max Planck Institute for Empirical Aesthetics; five behavioral studies with 235,440 judgments from US and South Korean participants show that timbre reshapes consonance preferences.

  5. September 2024

    Language models and sensory judgments

    In September 2024, Scientific Reports publishes "Large language models predict human sensory judgments across six modalities", with him as first author alongside Ilia Sucholutsky, Pol van Rijn, Nori Jacoby and Thomas Griffiths. Pairwise similarity judgments elicited from GPT models correlate with human data across six psychophysical datasets and recover the color wheel and the pitch spiral.

  6. November 2025

    A Kempner candidate talk

    In November 2025, he presents "Aspects of Similarity in Natural and Artificial Minds" from Princeton University as a Kempner Institute research fellow candidate. The talk uses AI, Bayesian modeling and human-in-the-loop paradigms to revisit classic questions about similarity, alignment and abstraction in humans and machines.

  7. 2026

    A Kempner Research Fellowship

    In 2026, he is named one of six Kempner Institute Research Fellows at Harvard, each serving a term of up to three years. He arrives with a psychology PhD from Princeton, advised by Tom Griffiths in the Griffiths Computational Cognitive Science Lab, and an MSc in physics and a joint BSc in physics and electrical engineering from the Technion.

  8. As of October 2026

    Research Fellow at Kempner

    He is a Research Fellow at the Kempner Institute at Harvard University, studying the computational principles underlying effective cognitive representations with AI, large-scale behavioral studies and "algorithms with people". Starting in academic year 2027–28, he is due to join the Department of Psychology at Tufts University as an Assistant Professor.

Career

  1. Since 2026Harvard UniversityResearch Fellow

Education

  1. TechnionBSc in Physics and Electrical Engineering
  2. Princeton UniversityPhD in Psychology

From the biography.

  1. More people shaping AI studied at Massachusetts Institute of Technology than at any other institutionIncludes Harvard University. We estimate where the people shaping AI studied and where their careers led. Massachusetts Institute of Technology leads with 467 former students, including 291 recorded doctorates.
  2. The median h-index is 48 at universities and 14 at frontier AI labsIncludes Harvard University. 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. Companies supply the most prior-work hires at 7 of the 9 frontier AI labsIncludes Harvard University. 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.
  4. 44% of AI career moves go from one university to anotherIncludes Harvard University. We estimate how 26,439 AI career moves link universities, companies, labs, nonprofits and government research centers, with decade counts, net flows and the top paths.

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 Raja Marjieh yet. Send us a free-licensed photo.

  1. Gibbs Sampling with People proceedings.neurips.cc. Checked 22 October 2026.
  2. Raja Marjieh raja-marjieh.github.io. Checked 22 October 2026.
  3. NeurIPS 2022 Awards neurips.cc. Checked 22 October 2026.
  4. Using Natural Language and Program Abstractions to Instill Human Inductive Biases in Machines arxiv.org. Checked 22 October 2026.
  5. Words are all you need? Language as an approximation for human similarity judgments (ICLR 2023) iclr.cc. Checked 22 October 2026.
  6. Words are all you need? Language as an approximation for human similarity judgments arxiv.org. Checked 22 October 2026.
  7. Timbral effects on consonance disentangle psychoacoustic mechanisms and suggest perceptual origins for musical scales nature.com. Checked 22 October 2026.
  8. Large language models predict human sensory judgments across six modalities nature.com. Checked 22 October 2026.
  9. Aspects of Similarity in Natural and Artificial Minds - Kempner Institute kempnerinstitute.harvard.edu. Checked 22 October 2026.
  10. Announcing 2026 Kempner Institute Research Fellows - Kempner Institute kempnerinstitute.harvard.edu. Checked 22 October 2026.
  11. Raja Marjieh - Kempner Institute kempnerinstitute.harvard.edu. Checked 22 October 2026.