Domino Data Lab at NVIDIA’s GTC announced powerful new updates giving every enterprise access to cutting-edge open-source tools and techniques to achieve AI value sooner.
Domino’s Spring 2023 release expands enterprise-grade support for open-source ML tools — Ray 2.0, MLflow, and Feast’s feature store — used to develop today’s most advanced AI apps. Additionally, Domino Cloud, a new fully-managed MLOps platform-as-a-service is now available for fast and easy data science at scale. Finally, Domino’s hybrid- and multi-cloud Nexus capability is now generally available with new fractional GPU options.
Accelerating Innovations Such as Generative AI with State-of-the-Art ML Tools
Domino now supports version 2.0 of the Ray open-source framework, which enables data science teams to rapidly develop and train generative AI models at scale, including ChatGPT. The integration with Domino’s on-demand, auto-scaling compute clusters streamlines the development process, while also supporting data preparation via Apache Spark and machine learning and deep learning via XGBoost, TensorFlow, and PyTorch.
“Ray is the ideal distributed processing option to scale machine learning, however it requires DevOps cycles to provision and manage a dedicated cluster for Ray jobs, which you pay for even while it sits idle between distributed training jobs,” said Till Buchacher, Head of Data Science at Direct Line Group. “With Domino, our users spin up Ray clusters on demand when needed. Domino handles the DevOps so they can focus on delivering quality work.”
Domino’s integration with MLflow simplifies machine learning lifecycle management for data scientists. It enables data science teams to track, reproduce, and share machine learning experiments and artifacts within their Domino projects, while Domino’s security layer ensures metrics, logs, and artifacts are secured.
Feast, an open-source feature store for machine learning, now integrates natively within Domino, providing users with easy access to query and transform ML features. This integration allows teams to reuse feature logic consistently and efficiently across data science projects, while tracking feature lineage and ensuring data accuracy and security, as well as cost savings from not re-computing business logic for each feature.
Accelerating Time to Value with a New Fully-Managed Enterprise MLOps Cloud Service
Domino also launched Domino Cloud, a fully-managed software-as-a-service version of its MLOps platform. It reduces AI time-to-value by providing scalable resources and a secure, governed enterprise-grade platform without any platform setup or management investment. Customers can save costs by paying only for the compute used while still accessing GPUs and distributed compute frameworks. Domino Cloud eliminates the need for data science teams to worry about deploying, upgrading, or managing infrastructure, allowing them to focus on their core responsibilities.
Accelerating AI Innovation Without Infrastructure Silos
Announced in June 2022 with NVIDIA as the first launch partner, Domino Nexus is now generally available to enterprises with powerful accelerated computing for workloads like generative AI across hybrid- and multi-cloud environments. A member of the NVIDIA AI Accelerated program, Domino workloads can be deployed from data centers to the edge, with seamless workload migration across cloud and on-premises environments.
“Now more than ever, companies realize that they must integrate AI into their business or be left behind,” said Nick Elprin, CEO and co-founder at Domino Data Lab. “ChatGPT has popularized the power of AI, but most organizations lack easy access to the tools and techniques to build and deploy such advanced models. Domino makes that power accessible to everyone, and we can’t wait to see what our customers will create.”
Domino also announced new cost- and performance-enhancing Nexus capabilities via a new validation with Vultr, a leading independent cloud computing platform. It enables Domino Nexus customers to seamlessly burst to Vultr Cloud with virtualized fractional NVIDIA A100 Tensor Core GPUs. The powerful combination of Domino’s Enterprise MLOps platform, Vultr infrastructure, the NVIDIA NGC catalog, and the NVIDIAÂ AI Enterprise software suite makes innovations in generative AI, computer vision, and more accessible and affordable for all enterprises. Data science teams can focus on delivering business impact, while IT has confidence in a validated solution architecture giving granular governance over cost, performance, and security.
“To capitalize on the power of generative AI models, enterprises are looking for AI solutions that integrate easily into their operations and deliver on performance and costs,” said Manuvir Das, Vice President of Enterprise Computing at NVIDIA. “NVIDIA’s collaboration with Domino Data Lab provides customers a powerful hybrid- and multi-cloud MLOps solution with the flexibility required to maximize productivity throughout the AI development and deployment lifecycle.”
Availability
Domino’s Spring 2023 release (Domino 5.5), Domino Cloud and Domino Nexus are all available today. MLflow experiment tracking and Feast feature store integration are available in preview. The new Domino and Vultr solution will be available later this year.
For more information, visit the website here.
Related News:
