AI-Powered R&D Platform


An Efficient End-to-end AI Solution for Early Drug Discovery.

AIxMolTM is created jointly by a cross-discipline in-house team of scientists and engineers including CS engineers, CADD and AIDD specialists and medicinal chemists with the goal to discover the ideal drug candidates much more efficiently than the traditional methods. It is an integrated platform for drug discovery and optimization, providing the essential tools such as molecular generation, virtual screening, ADMET prediction, and binding free energy calculation, etc. to enable speed and precision. The platform also includes a series of specialized AI tools to enable product differentiation for our own pipeline development such as tissue-selectivity model, multi-target drug design, etc.

Featured Tools

AI Model for Tissue Selectivity

Machine learning-based prediction and optimization of tissue distribution-related pharmacokinetic properties

Multi-Target Ligand Design and Screening

Molecule generation and screening platform based on binding site’s feature fusion of multiple targets

Basic Tools

Molecule Generation


A scaffold-hopping tool powered by deep learning

Virtual Screening + Active Learning

A Graph Convolutional Network (GCN) based scoring function for protein-ligand binding affinity prediction


A compound recommender system of multiple targets


A multi-task Graph Convolutional Network (GCN) model to predict the compound-protein interactions

Active Learning augmented GNINA and VINA

Integrated GNINA and VINA with Active Learning framework for ultra-large scale virtual screening

ADMET Predictor


A deep learning model co-developed with Baidu for the prediction of ADMET properties

Binding Free Energy Calculation

Relative/Absolute FEP

A calculation tool for binding free energy between small molecules and protein based on free energy perturbation theory