At MMLab, we e merge computational chemistry, machine learning, and biomimicry to create nature-learned molecules and materials. We utilise Density Functional Theory (DFT), Molecular Dynamics (MD) simulations, and coarse-grained (CG) models to simulate materials from the nanoscale to the mesoscale, covering chemical reactivity, electron transport, self-assembly, and degradation. Across these areas, we develop more efficient molecular representations for training machine learning (ML) models to accelerate property prediction and molecular discovery. Our philosophy is to “make it simple but significant,” creating effective models for complex systems.