We start with proprietary foundational ML models, trained on our platform data, not only public datasets. These models screen millions of candidate compounds and predict which are most likely to be active against specific skin biology targets.

Predicted hits move into automated cell-based assays. Robotic liquid handling and high-content microscopy let us test at scale with higher reproducibility than manual methods. This stage filters thousands of candidates down to the most promising.

This is where our platform is fundamentally different. We test in full-thickness human skin — epidermis and dermis with native architecture — maintained for four weeks. Compounds that work in a petri dish often fail here. Compounds that pass here have been validated in long-lasting, full-thickness native human tissue.

Validated compounds go through deep multi-omic characterization — transcriptomics, proteomics, secretomics, and phenotypic imaging. We map molecular activity to observable skin outcomes to understand not just that a compound works, but how it works and who it works for.

Compounds that clear our validation bar move into manufacturing scale-up, regulatory and commercial development. We develop ingredients that are ready for formulation — with mechanism data, efficacy evidence, and safety profiles that brands can build real claims around.