Atomistic intelligence from Liechtenstein
We build machine‑learned interatomic potentials (MLIP) and ML‑accelerated DFT to make materials modelling faster and more reliable.
- MLIP for large‑scale MD and materials screening
- ML‑accelerated DFT: surrogates, smart initialisation, and convergence aids
- Active‑learning loops with uncertainty quantification (UQ)
Early stage — private pilots with research partners.
e3nn-mlx is WORKING now!
Email: contact [] atom.li