Following the introduction of the Model Management framework at the Rev summit for data science leaders earlier this year, we’re delivering on our promise to provide a platform that allows organizations to put predictive and machine learning models at the heart of their business,” said Nick Elprin, co-founder and CEO at Domino. “Domino 3.0 tackles model operations challenges surrounding model delivery and ongoing iterations of production models. By streamlining the model deployment process and facilitating faster, easier iterations throughout the model management cycle, the latest Domino functionality helps data science leaders ensure that their investments in data science are yielding tangible business impact.”
The Last Mile Problem in Data Science
Despite the massive investment in artificial intelligence (AI) and data science, most organizations have been unable to see significant return. A recent BCG and MIT Sloan survey found that nearly 85 percent of the executives surveyed believe AI will offer their organization a competitive advantage, yet only five percent of companies extensively utilize AI models in their business.
Many data science investments fail because it’s extremely hard for companies to get models beyond their data science teams and into production,” stated Krishna Roy, senior analyst, Data Science & Analytics at 451 Research. “Organizations will continue to struggle to deliver business impact with data science if they don’t close the gap that exists between IT and DevOps teams responsible for data science production environments, data scientists building models, and end users and other business stakeholders involved in model creation and use.”
Domino 3.0 Helps Companies Become Model-Driven With LaunchPad
Domino 3.0 fills the gaps between AI investment, execution, and competitive advantage by eliminating the model delivery bottleneck.
Domino’s new module, Domino Launchpad, enables data science teams to decrease the time and pain associated with deploying models while increasing iteration speed on models in production to ensure they have an immediate business impact by:
- Eliminating technology hurdles to model delivery with automatic infrastructure provisioning via Docker and support for popular tools like Shiny, Flask and Dash;
- Improving collaboration among stakeholders with a single portal for discovering model products and usage data to ensure faster feedback loops and validate a model’s impact; and
- Accelerating model iteration velocity with automatic model versioning and full reproducibility of experimental history for faster, continuous model iteration.
As a leader supporting Global 2000 clients in the competitive media landscape, a core differentiator has become our ability to rapidly build and deploy client-facing AI apps at scale,” said Giovanni Romero, partner and global lead for Business Consulting and Analytics at GroupM/WPP’s Mindshare. “With data scientists and business stakeholders spread across more than 110 offices, we needed a more systematic way to incorporate feedback, ensure model reproducibility, and manage production models. Domino helps us understand and respond to clients’ engagement with AI apps so we can constantly improve the effectiveness of their media investments; the enhancements provided in Domino 3.0 streamline the workflows between data scientists and end users even further.”
Sign up for the free insideBIGDATA newsletter.