Predictive chemistry
We aim to improve the quality, interpretability, and operational performance of chemical predictions so models can support confident scientific decisions.
Explainable QSAR
We develop QSAR workflows that expose the drivers behind predictions, helping teams understand when and why a model should be trusted.
Improved QSAR accuracy
We develop advanced model architectures that improve QSAR accuracy, strengthen activity-cliff prediction, and incorporate proprietary contextualisation to support more reliable compound-level decisions.
Abstract chemical transformation engine
We are designing a next-generation chemical transformation engine to model abstract reaction logic and power applications in generative chemistry and metabolism prediction.