Our services

We provide comprehensive support for materials design driven by state-of-the-art machine learning methodologies.
Our service portfolio is engineered to deliver high-performance, quantum-accurate simulation capabilities.

Custom potential development

We specialize in the development of custom-made (bespoke) ACE and GRACE machine learning interatomic potentials. This process is conducted in close, continuous collaboration with our customers to ensure that the resulting models are precisely optimized for your specific material systems and unique challenges.

Foundational model optimization

To maximize efficiency and leverage existing data, we offer fine-tuning services for foundational (universal) models. We adapt these robust, general-purpose potentials for highly specific applications and local environments, ensuring quantum accuracy where you need it most.

Atomistic simulation and property prediction

We execute complex atomistic simulations to predict critical material properties that directly inform design cycles. Our capabilities include but are not limited to predictions of

  • Phase stability assessment
  • Phase transformations
  • Diffusion and transport
  • Thermal conductivity
  • Segregation and intermixing at interfaces and boundaries
  • Defect properties
  • Molecular conformations
  • Surface chemistry and reactions
  • Magnetic materials
  • Solvents and solutions
  • Amorphous materials

Technical support

We offer dedicated support for running and applying atomistic simulations within your existing infrastructure. This includes guidance on setting up simulations, selecting optimal parameters, interpreting results, and integrating our models into your computational workflows.

Software support

We provide full technical support for our software solutions, specifically PACEmaker and GRACEmaker. This ensures your team can seamlessly implement, run, and maintain our Performant Atomic Cluster Expansion (PACE) and Graph Atomic Cluster Expansion (GRACE) models, maximizing your simulation throughput and reliability.

Software development

We develop custom-made software solutions for atomistic simulations, from tools for efficient atomistic simulation workflows to complete, dedicated software for efficient training of ACE and GRACE machine learning interatomic potentials.