Our technology has been published in leading peer-reviewed journals and validated on projects with the world's largest pharmaceutical companies.
We pioneered a class of models for molecular properties prediction that know what they know as well as know what they don't know.
We designed a new algorithm that accurately predicts bioactivity in the low-data limit, validated in collaboration with Pfizer to achieve 2x state-of-the-art performance in prospective experiments.
We designed the first algorithm to outperform human chemists in predicting the outcomes of chemical reactions, tackling challenges in chemo-, regio- and stereo-selectivity.
Collaborating with Pfizer, we showed that our reaction prediction algorithm is generalizable to pharma chemical space, and we can predict synthetic routes to complex drug-like molecules.
Let us show you how our technology can help.
We pioneered a class of models for molecular properties prediction that know what they know as well as know what they don't know.
We designed a new algorithm that accurately predicts bioactivity in the low-data limit, validated in collaboration with Pfizer to achieve 2x state-of-the-art performance in prospective experiments.
We designed the first algorithm to outperform human chemists in predicting the outcomes of chemical reactions, tackling challenges in chemo-, regio- and stereo-selectivity.
Collaborating with Pfizer, we showed that our reaction prediction algorithm is generalizable to pharma chemical space, and we can predict synthetic routes to complex drug-like molecules.
Let us show you how our technology can help.
