Alireza Ghafarollahi · Postdoctoral Researcher · MIT

Accelerating scientific discovery through AI.

I develop multi-agent AI frameworks that integrate large language models with physics-based simulation for autonomous materials design and discovery. This work builds on a foundation in statistical mechanics, atomistic simulation, and the theory of defect kinetics in structural alloys.

Alireza Ghafarollahi
Published in
PNAS Advanced Materials Acta Materialia Science Advances MRS Bulletin npj Computational Materials Digital Discovery
0Peer-reviewed publications
0Citations (Google Scholar)
01 / About

From atoms to autonomous agents

My research addresses a central challenge in computational science: how to couple artificial intelligence with rigorous physical theory to accelerate the discovery and design of materials. At MIT, working with Prof. Markus Buehler, I develop LLM-driven multi-agent frameworks — Sparks, SciAgents, AtomAgents, and ProtAgents — in which specialized agents orchestrate molecular dynamics and density functional theory calculations, machine-learned interatomic potentials, and graph neural networks to autonomously generate hypotheses, execute simulations, and interpret results across materials and molecular systems.

My doctoral research at EPFL with Prof. William Curtin developed parameter-free, transition-state-theory descriptions of thermally activated dislocation processes — double-kink nucleation and kink migration — in BCC dilute and high-entropy alloys. By computing activation free-energy barriers for these rare events and validating the theory against large-scale molecular dynamics, this work established predictive, atomistically informed models of temperature- and strain-rate-dependent strength.

Together, these two lines of work define my research program: physics-grounded autonomous AI systems in which statistical mechanics, atomistic simulation, and machine learning operate as an integrated engine for scientific discovery.

2023 — presentPostdoctoral Researcher, Massachusetts Institute of TechnologyAgentic AI for scientific discovery · Cambridge, USA
2022 — 2023Postdoctoral Researcher, Max Planck Institute for Sustainable MaterialsAtomistic study of fracture mechanisms in defected alloys · Düsseldorf, Germany
2018 — 2022Ph.D. Mechanical Engineering, EPFLStatistical mechanics of dislocations in BCC & high-entropy alloys · Lausanne, Switzerland
2012 — 2014M.S. Civil Engineering, Sharif University of TechnologyElastic wave scattering & micromechanics · Tehran, Iran
2008 — 2012B.S. Civil Engineering, Amirkabir University of TechnologyTehran, Iran
02 / Recognition

Awards & media

🏅
MIT CEE Postdoctoral Scholar Mentoring, Teaching and Excellence AwardMassachusetts Institute of Technology · 2026
🎓
SNSF Postdoc.Mobility FellowshipSwiss National Science Foundation · 2022
03 / Research

Research programs

Two complementary research thrusts: autonomous AI systems for scientific discovery, and predictive statistical-mechanics models of alloy strength.

Flagship AI systems

Physics-aware multi-agent frameworks developed at MIT that execute the research cycle end to end — hypothesis generation, simulation, and analysis.

04 / Publications

Peer-reviewed publications

20 papers across leading materials, physics, and AI venues — including PNAS, Advanced Materials, and Acta Materialia.

Citations per year

1,230+ total · Google Scholar ↗
05 / Teaching & Mentoring

In the classroom

Teaching Assistant — MIT

  • Atomistic Modeling and Simulation in Materials and Structures · Fall 2024
  • Advancing Mechanics and Materials via Machine Learning · Spring 2025

Teaching Assistant — EPFL

  • Advanced Solid Mechanics · Fall 2018, 2019, 2020
  • Introduction to Structural Mechanics · Spring 2019, 2020, 2021

Research Mentoring — MIT

  • Mentoring an MIT EECS undergraduate on machine learning prediction of solute/screw interaction energies in BCC alloys

Peer Review

  • Nature Machine Intelligence · Journal of the Mechanics and Physics of Solids · MRS Bulletin · Digital Discovery · Journal of the Mechanical Behavior of Biomedical Materials
06 / Contact

Get in touch

I am currently on the academic job market, seeking faculty positions in computational science, materials science and engineering, and AI for science. I welcome inquiries regarding open positions, research collaborations, and speaking engagements.