CDI Lab is enabling flexible and cost-effective access to cloud-based computational resources which can be used in AI (model development and training), scientific computing, and heavy simulations. This is suited to organisations that do not have their own hardware resources or cloud infrastructure management skills, yet require cost efficiency and controlled access to sensitive information.
CDI Lab enables researchers to work in their favourite environments (JupyterLab, VSCode, CLI tools), and connect these with AWS cloud infrastructure allocated on demand and automatically released when no longer needed.
Researchers log into CDI Lab via their accounts, which could be linked to institutional single-sign-on, to ensure user access control and budgeting by the organisations.
A large number of packages are pre-installed, and users can further install new packages. Working environments are containerised and self-repairing.
CDI Lab spins servers as requested by the user. The shared server is for general code development. This resource is used by everyone from the organisation. For model training researchers may choose a standalone server, selecting server size from a drop-down menu. For large jobs such as AI/ML model training or complex data pipelines a Ray cluster is suggested, which will auto-scale to the required number of servers to match the workload demands.
If you’re interested in booking a demo or would like to learn more about pricing, please contact us at info@cdi-sg.com