MathWorks has been accelerating data science and engineering since 1984. Today, their MATLAB® and Simulink® products are some of the world’s easiest and most productive engineering and scientific design environments with powerful algorithms and domain specific models. Its community of more than four million users are actively incorporating models into workflows across verticals such as automotive, aerospace, and energy to improve productivity and shift focus from mundane work to value-building intellectual efforts.
Challenges in Becoming a Model-Driven Business
Integration of models into the heart of products and business processes will drive unprecedented growth over the next decade, enabling model-driven businesses to upend competitors with more accurate predictions. In the years since companies began experimenting and slowly moving models into production scenarios, many recurring challenges have emerged:
With multiple data science groups choosing different technology stacks and data science languages (like MATLAB, Python® and R), sharing knowledge across teams is limited by platform disparity and lack of integrations.
Complex data science work requires access to powerful tools and compute resources. A workstation with the necessary processing power for tasks such as deep learning gets expensive quickly - a multi-GPU desktop can run into the tens of thousands of dollars.
Regulators in the financial and healthcare industries need to understand how models were developed, and what data and code were used to create them. This requires tight controls on IT governance, security, and reproducibility.
Domino + MathWorks: Enabling Scalable Infrastructure and Shared Knowledge to Accelerate Model Development
Solving these very challenges for joint customers is what brought Domino Data Lab and MathWorks together. Our partnership announcement this week at MATLAB EXPO builds on the success of working to deliver best-in-class integrations from both companies in a web-first experience.
Our companies are aligned in the mission to accelerate the development and deployment of data science work while increasing collaboration and governance. For MathWorks’ MATLAB® and Simulink® users, the benefits of the Domino partnership are clear.
Knowledge sharing and broad support of tools allows data science teams to build on past work and freely collaborate with their peers:
- Easy access to MATLAB from anywhere — including all major cloud platforms. Users access their tools through a browser, sparing the need to invest in expensive workstations that only get utilized during work hours.
- Shared datasets in Domino make discovery of data easier and allows teams to collaborate with multiple modelling efforts.
- Improved collaboration with users of other tools (JupyterLab, RStudio, SAS, etc.) to work together to solve the most difficult business challenges.
- Automatic versioning so projects in MATLAB and Simulink can be reproduced at any time to support audits and compliance requirements.
Agile infrastructure and self-service sandboxes give data science teams power and flexibility without sacrificing IT governance.
- Democratized access to NVIDIA GPUs and other powerful compute (e.g. more CPU cores, large RAM machines, and distributed workloads on frameworks such as Spark and Ray) accelerates ML-related tasks, from labeling data to training models
- More efficient simulation and modeling in Simulink and Simscape thanks to the ability to stand up a container with dozens of CPU cores.
- Seamless integration with Domino’s batch execution for scheduled training and launchers to share MATLAB and Simulink algorithms with peers.
Best of all, MathWorks’ biannual releases can be rolled out instantaneously across the organization without a need for installation across the whole enterprise’s desktop fleet. Adding new MATLAB release images to the Domino platform allows users to switch seamlessly between versions allowing for validation of existing models with updated versions. .
Real-World Impact with Market-Moving Potential
Our partnership is still in the early stages, but we are already seeing enthusiasm in the market. Customers appreciate these shared capabilities overall, and specifically: getting answers fast. IT and data science leaders appreciate the hardware and support cost savings. At the same time, Domino’s project tracking tools provide them with clear insight into project progress and expenditures.
Lockheed Martin has unlocked more than $20m in data science efficiencies and cost-savings across 300 data scientists by centralizing tooling across the enterprise with Domino and MathWorks, with NVIDIA as its choice in high-powered computing. With this three-pronged solution, it streamlined collaboration and knowledge sharing while automating manual DevOps tasks that had hindered data scientist’ productivity.
A global pharmaceutical company has accelerated its fight against cancer and other chronic and life-threatening conditions, successfully testing tens of thousands of hypotheses with complete reproducibility while accelerating the submission of test results to the FDA. It was able to break down silos across 500-plus data scientists working in MATLAB, Python, R, and SAS, as well as productionize models more quickly to publish findings to cross-functional “consumers” of model output - all in the Domino platform.
We look forward to enabling additional features and helping customers drive model-driven innovation.