Scientist · Entrepreneur
I build infrastructure for better science.
I am a scientist and entrepreneur. I work at the intersection of materials science, software, and artificial intelligence—turning state-of-the-art research into reliable software and real-world impactful outcomes.
Current thesis
The next leap in science will come not only from better models, but from better systems for turning knowledge into action.
That means reproducible workflows, AI-ready data, clearer tools, and organizations designed to learn quickly. When the infrastructure is right, scientists spend less time managing complexity and more time making discoveries.
Latest thinking
Why materials R&D is shifting toward productization
The difference between a project and a product is compounding. A product gets better with every use. A project is reinvented every time.
Everything around us is made of materials, yet we rarely think about it
That is part of why I find this field so meaningful: small improvements in materials quietly reshape what is possible in energy, health, and technology.
WTH is Materials Informatics??
Meta, Google, and Microsoft all made news in materials informatics in 2023. What the field actually is, and why it is suddenly crowded.
Cloud security for R&D
The question every research organisation asks before moving mission-critical work off its own hardware, and what the answers looked like in practice.
What I’ve built
Mat3ra
AI-driven materials science in the cloud: design a structure, run the simulation, and build models on the results, without standing up HPC infrastructure first.
Density-functional theory, molecular dynamics, and machine-learned force fields run through one workflow engine and API, wrapping VASP, Quantum ESPRESSO, LAMMPS, and GROMACS. Used by Shell, the U.S. Air Force, the U.S. Department of Energy, Tokyo University of Science, and Tohoku University.
Open scientific tools
The libraries underneath the platform, released openly so the data outlives any one vendor—including ours.
Made for creating and manipulating structures, wave.js for atomic visualisation in the browser, ESSE for the JSON schemas describing materials and workflows, and Express for extracting properties from simulation output.
Research
AI-ready design of realistic 2D materials and interfaces with Mat3ra-2D
An open-source framework for generating two-dimensional materials and interfaces with realistic disorder and defects, written with a co-author at NIST.
Interpretable machine learning for materials design
On the trade-off between how well a model predicts a material property and whether a scientist can understand why, tested across four machine-learning methods.
Data-centric online ecosystem for digital materials science
The argument underneath the platform: materials data is only useful if it is standardised, shareable, and outlives the project that produced it.
Superconductivity and electron-phonon coupling in lithium at high pressures
Doctoral work at Berkeley with Jesse Noffsinger and Marvin L. Cohen, on why lithium becomes a superconductor when it is squeezed.
Leadership
AI at the Frontier
Spoke on how cloud computing and artificial intelligence are accelerating the discovery of materials behind batteries, chips, and aerospace, at Deep Tech Week in San Francisco.
Artificial Intelligence for Materials Science
Presented “Data standards: the key enabler of AI-driven materials” and joined the panel on the challenges and prospects of artificial intelligence in the field.
Consortium for the Advancement of Materials Science with AI
Helped establish a consortium of companies, universities, and research institutions advancing open innovation in artificial intelligence for materials and chemistry.
Founded the materials R&D cloud
Started the company, then Exabyte Inc., out of Alchemist Accelerator, and lead it today: a platform used across academia and industry to design materials and run simulations at scale.
About
Born in Kazan, Tatarstan, I trained as a physicist at MIPT and UC Berkeley, where my doctorate went into first-principles studies of superconductivity and beyond.
That work ran on the simulation codes I have spent the years since trying to make easier to live with. The physics-based materials modeling engines - exact, powerful, are unreasonably hard to use: an hour spent fighting a supercomputer or rebuilding someone else’s environment is an hour not spent on advancing science. I founded Mat3ra.com to close that gap.
Here I write about science, entrepreneurship, and building systems that let smart people accomplish their work faster.
- Current work
- Founder and chief executive officer, Mat3ra
- Education
- Physics doctorate, UC Berkeley,
with Marvin L. Cohen - Research
- 20+ papers, h-index 15
- Backing
- Breakout Labs, then a seed round led by
Tim Draper, with Albert Wenger and Serg Bell
Beyond the work
The human side matters, too.
Outside work, I spend time on a bike—often around Mount Diablo. Long climbs are a useful counterweight to abstract problems: physical, direct, and a reminder that progress is usually accumulated one steady effort at a time.
Let’s work on something that matters.
I am always interested in thoughtful conversations about science, technology, entrepreneurship, and building organizations for the long term. If you are a scientist or engineer who would rather build this than read about it, Mat3ra is hiring.