A Fundamental Shift In Drug Discovery Driven By AI
We've put technology to work to accelerate drug discovery in ways never possible before. And now we are democratising access to AI through our flagship SaaS platform ROSALINDᴬᴵ
The world's largest repository of state-of-the-art AI models
Off-the-shelf models to accelerate drug discovery
At the core of Rosalindᴬᴵ is a large suite of 100s of ML models pretrained for de novo design of chemistry and accurate prediction of 50+ chemical properties. These models can be used to inform novel chemistry and prioritisation of compounds for lab testing.
Intuitive interface for custom-built ML pipelines
Rosalindᴬᴵ helps drug hunters train best-in-class ML on their own internal catalogues. Using our intuitive interface, you can accelerate progress on multiple programs by training, analysing and driving insights that help you make the best decisions.
Experience the future of AI for drug discovery
Our platform can generate insights from a fraction of the data required by other computational methods.
We are building best-in-class ML models from as little as 10% of the lab data usually required.
Our methods learn from every data sample and ensure the best possible performance at a significant reduction in lab testing.
The ratio of positive to negative samples in drug discovery greatly impacts ML performance.
We've built ML models that can be trained on severely unbalanced with only as little as 0.0001 of the data is positive samples.
Our methods learn generalisable patterns from the very small set of positive samples and we use these patterns to discover hidden insights.
Lab experiments are prone to errors and noise that make training accurate ML models challenging. We have invented a new method to learn from noisy data while achieving state-of-the-art performance.
Our methods work on classification and regression tasks and support finding, quantifying, and learning with label errors in data.
All you need to drive your R&D at the cutting edge of AI
With ROSALINDᴬᴵ, you get access to a large collection of pretrained cutting-edge ML models trained on a large catalogue of libraries providing a 360° view of chemical properties
ROSALINDᴬᴵ pretrained models can be easily refined on your internal libraries. This helps bootstrap your training processes while ensuring best results for your organisation
ROSALINDᴬᴵ automates the ML pipeline from data curation to model optimisation and reporting. You get access to traditional and advanced deep learning models without the development hassle
Share your models built on ROSALINDᴬᴵ with your colleagues and collaborators to help accelerate drug discovery programs and drive innovation within and across organisations
How ROSALINDᴬᴵ supports drug hunters
Instant access to a large collection of public and proprietary datasets including compounds and activity data
Comprehensive chemical descriptors to represent chemical and biological data for ML pipelines
Scalable cloud infrastructure to support large scale training of ML and heavy physics-based calculations
Orchestration of automated complex data pipeline to clean, normalise and prepare data for ML
Accurate activity prediction to accelerate data-driven decision making and highlight hidden insights with a secure and scalable ML
Secure and govern platform data and metadata and control capabilities
A comprehensive cloud-based Edge-to-AI analytics services to visualise and inspect your data
De novo Design
Access uncharted chemical spaces relevant for specific targets with cutting-edge generative AI
Fine-tune your workflow independently to better utilize your data and internal resources