Boris Kozinsky
Boris Kozinsky is a materials scientist whose machine-learning methods extend atomistic materials simulations to large systems. After studying physics, mathematics, and electrical engineering and computer science at MIT, he earned a physics PhD in 2007 and founded the Bosch RTC Research Center in Cambridge. In 2018, he started a Harvard group; his research collaborations produced FLARE, NequIP, Allegro and Allegro-pol for atomistic simulations and electric-field response. He became an APS Fellow in 2023, received Harvard's McDonald mentoring award in 2024, and is Gordon McKay Professor of Materials Science and Mechanical Engineering and Professor of Chemistry and Chemical Biology at Harvard as of October 2026.
- Now at
- Harvard University
- Works
- 281
- Citations
- 12,570
- h-index
- 49
- Badges
Publication numbers from OpenAlex, as of 27 October 2026.
Timeline
-
By 2007
MIT degrees and Bosch RTC
He earns bachelor's degrees in physics, mathematics, and electrical engineering and computer science at MIT, then earns a PhD in physics there in 2007. In 2007, he founds the Bosch RTC Research Center in Cambridge, Massachusetts.
-
2018
Harvard research group
In 2018, he starts a Harvard group that combines physics and machine-learning models to study transport properties and the time evolution of complex materials. He connects the Harvard and Bosch teams through close collaborations.
-
2020
FLARE and active learning
In 2020, he coauthors FLARE, a Bayesian active-learning method that builds interpretable interatomic force fields from molecular-dynamics structures and uses model uncertainty to decide when to request first-principles calculations.
-
2022
NequIP interatomic potentials
In 2022, he coauthors NequIP, an E(3)-equivariant neural network for interatomic potentials that models atomic geometry and reports state-of-the-art accuracy with up to three orders of magnitude fewer training data than existing models.
-
2023
Allegro at large scale
In 2023, he coauthors Allegro, a strictly local equivariant neural network for interatomic potentials. The paper demonstrates parallelization with a simulation of 100 million atoms.
-
2023
An APS Fellowship
In 2023, he becomes a Fellow of the American Physical Society for developing computational and machine-learning methods to study microscopic transport and dynamic phenomena in materials for energy storage and conversion.
-
2024
A Harvard mentoring award
In 2024, he receives the Capers W. McDonald and Marion K. McDonald Award for Excellence in Mentoring and Advising at Harvard SEAS.
-
2025
Allegro-pol electric response
In 2025, he coauthors Allegro-pol, an equivariant machine-learning framework for predicting electric-field responses under physical constraints. The model predicts material responses with quantum-level accuracy for systems up to a million atoms.
-
As of October 2026
Harvard professor and MRSEC lead
As of October 2026, he is the Gordon McKay Professor of Materials Science and Mechanical Engineering and Professor of Chemistry and Chemical Biology at Harvard. For 2026–2027, he leads MRSEC IRG I, Designing and Controlling Phase Transitions in Functional Molecular Materials.
Career
- NowHarvard UniversityGordon McKay Professor of Materials Science and Mechanical Engineering, Professor of Chemistry and Chemical Biology
- From 2007Bosch RTC Research CenterFounder of the Bosch RTC Research Center
Education
- 2007Massachusetts Institute of TechnologyPhD in physics
- Massachusetts Institute of TechnologyBachelor's degrees in physics, mathematics, and electrical engineering and computer science
Colleagues
Sources and credits
Photo: We have no published photo of Boris Kozinsky yet. Send us a free-licensed photo.
- Kozinsky receives McDonald Mentoring Award
- Prof. Boris Kozinsky
- On-the-fly active learning of interpretable Bayesian force fields for atomistic rare events
- E(3)-equivariant graph neural networks for data-efficient and accurate interatomic potentials
- Learning local equivariant representations for large-scale atomistic dynamics
- Kozinsky named American Physical Society Fellow
- Modeling electric response of materials, a million atoms at a time
- Unified differentiable learning of electric response
- Boris Kozinsky | Harvard John A. Paulson School of Engineering and Applied Sciences
- Faculty Participants | MRSEC at Harvard