AICE Lab

Atomistic Intelligence for Catalysis and Energy

Atomistic simulations and AI for energy materials and catalysis.
We develop ML-driven simulation methods to understand how materials behave under real operating conditions.

Core Philosophy

Artificial Intelligence and Atomic Simulation for Catalysis and Energy, led by Dr. Seokhyun Choung. Real catalysts operate at dynamic interfaces, where surfaces restructure mid-reaction and particles reorganize under working conditions. We tackle these problems with AI methodology: generative models, agents, and large language models, coupled with atomistic simulation to solve the dynamic-interface problems of catalysis and energy.

Experiment as Ground Truth

"No matter how beautiful the theory is."

- R. Feynman (1964)
Adopt AI Responsibly

Take responsibility for the quality of AI-generated outcomes.

Research Directions

AI Multi-Agent System for Material Discovery

Multi-agent systems that run the discovery cycle end to end: propose, simulate, analyze, and decide the next experiment. LLMs, generative models, and simulation tools orchestrated into autonomous workflows.

ICLR 2026 AI4Mat
Chem. Eng. J. (2024)
ACS Energy Lett. (2025)

MLIP AI Infrastructure for Material Discovery

Foundation-model potentials adapted to target systems with minimal DFT data, compressed into fast and accurate MLIPs for large-scale simulation.

AI4Mat-NeurIPS (2025)
Cell Rep. Phys. Sci. (2025)
Nat. Sensors (2026)

Operando Simulation for Real-world Impact in Collaboration with Experiment

MLIP molecular dynamics at realistic operating conditions, compared directly with in-situ experiments (XAS, DRIFT, TEM). 17+ experiment-theory co-publications.

Nat. Commun. (2026)
Nat. Commun. (2026)
Appl. Catal. B (2026)
What We Offer
AICE Lab Starter Pack - Mac mini, MacBook, DJI mic, Claude Max, gym membership
For Your Research
  • Claude Max for every lab member
  • GPU clusters (A6000/L40S, H100 via KISTI)
  • VASP, LAMMPS, PyTorch, ASE, custom tools
  • Direct experimental collaborator connections
For Your Life
  • Gym membership (we pay for it)
  • Flexible working hours (output > hours)
  • Unlimited deep discussions with Dr. Choung
  • 1-on-1 mentoring for YOUR career goals
  • Asking "why?" is encouraged, not punished
FAQ
Do I need prior DFT or ML experience?
No. We'll teach you everything. We care about curiosity and drive, not a perfect GPA or prior experience.
I only have experimental experience. Can I apply?
Yes, and we actually welcome it. Computational researchers who understand experiments run the best simulations. We'll teach the coding.
Is the lab English-friendly?
Yes. All group meetings and internal communication can be in English. Papers are in English. International students are welcome.
Do you really provide Claude Max?
Yes. AI tools are research infrastructure. We invest, not economize.
How many hours per week?
We don't count hours. Ask good questions, make steady progress, and take care of your health. That's it.
What's the lab culture like?
We value: intellectual honesty, curiosity, kindness, and taking care of yourself. We don't value: performative busyness, hierarchy for its own sake, or suffering as a badge of honor.
How are research topics decided?
Initially the advisor suggests directions, then you gradually take the lead. The ultimate goal is for you to find your own compelling questions.
Career paths after graduation?
Academia, national labs, industry R&D, AI/ML engineering, and more. The combination of atomistic simulation + AI skills is in demand everywhere.
How much coding skill do I need?
Basic Python is enough. You'll learn the rest on the job. AI tools have lowered the barrier significantly.

Interested?

Graduate students, postdocs, and undergrads all welcome. No simulation experience required.

No formal deadline. Positions open until filled.