Skip to content

    Each video highlights how an important data science task is improved using Domino. See how you can eliminate infrastructure friction, foster collaboration, and improve model velocity.

    Demo 1: Data Science Development

    Get an introduction to the powerful workbench capabilities in Domino that give data scientists self-serve access to the tools they want to use and the scalable compute they need to accomplish the most challenging data science projects – without DevOps complexity.


    You’ll see how to:

    • Start, stop, and change a Durable Workspace that provides easy access to a variety of open source and commercial tools, including JupyterLab, RStudio, SAS, and MATLAB.
    • Add scalable compute to match the development job, including GPUs and distributed compute (e.g. Spark, Ray, and Dask).
    • Check-in code with your repository of choice.
    • Share inputs and outputs from different code libraries, such as feeding inputs from a SAS model into a Python model.

     

    Demo 2: Collaboration and Project Management

    See how Domino enables teams of data scientists using different tools to seamlessly collaborate on a project, with the ability to leverage valuable insights, reproduce prior work, and harvest collective wisdom.

    You’ll see how to:

    • Use knowledge management features such as tagging, comments, and deep linking to organize projects more effectively.
    • Review activities to see all work being conducted on a project to understand the context.
    • Search prior work to eliminate “reinventing the wheel”.
    • Fork projects into new projects, or mark them as reference projects for code snippets in repeatable/reusable templates.
    • Manage your team’s projects more efficiently, with clear goals/metrics.
    • Reproduce work from any point in the past, with integrated version control and artifact tracking.

     

    Demo 3: Model Publishing

    Learn the four different ways to turn models into end-user assets that can be used by anyone in your organization to make better decisions and transform the way your business operates.

    You’ll see how to:

    • Deploy models as APIs that can be integrated into existing software applications and websites to enable prediction-on-demand services.
    • Publish interactive web applications and dashboards for executives and other business stakeholders.
    • Create launchers as parameter-driven wizards to support ‘what-if’ analysis and re-run particular parts of model code.
    • Schedule jobs for model scoring and write results into databases or tune/re-train models.
    • View all model assets that have been published to analyze usage and other useful statistics.