Case Study
Investment firm achieves breakthrough efficiency: 65% faster analytics with UST Xpresso
OUR CLIENT
This European multinational company provides investment management solutions to clients worldwide. With more than £200 billion in assets under management, it employs nearly 1,000 people.
THE CHALLENGE
Maintaining a diverse analytics environment
The data scientists at this investment company analyze historical asset performance data, market trends, and other key investment factors to help asset managers make data-driven decisions to optimize investment strategies and asset portfolios. Each data scientist used the analytics tools they liked best. The different applications were complex and challenging for the IT team to maintain. The company wanted to implement a standardized analytics infrastructure that would be easier to manage and provide the highly skilled data scientists with robust tools to streamline and accelerate their work.
THE TRANSFORMATION
UST Xpresso MLOps platform consolidated and streamlined analytics capabilities
After conducting a thorough analysis of the existing IT environment and analytics tools as well as the needs of the data scientists, UST Xpresso was deployed. Its machine learning operations (MLOps) platform enabled the analytics team to build, train, and deploy AI, ML, and large language models (LLMs) efficiently at scale. The solution was designed so the data scientists can easily:
- Access the company’s Python libraries.
- Create analytical pipelines using modular components with built-in design and coding best practices.
- Experiment and compare results for different analytical models.
- Test ML models on different hardware, such as CPUs and GPUs.
- Deploy final models using simplified, abstracted processes in the company’s IT environment.
- Schedule analytical pipelines to run at specific intervals.
- Manage governance mechanisms throughout the development, staging, and production lifecycle.
- Use a sandbox to transfer data and explore different developer tools, environments, hosted applications, and code repositories.
Now, the data science team can automatically pull and merge data from various in-house and external data sources, create and test data pipelines on the IT infrastructure of their choice, and easily deploy ML models into production.
Since the solution was deployed as a managed service, the company’s IT team doesn’t have to maintain the application or underlying infrastructure, allowing the team to focus on more strategic projects. The data scientists and IT personnel can contact the UST Xpresso team for help whenever necessary. In a recent support effort, the UST Xpresso team helped the company provision new hardware in just a few days, saving the IT team weeks of work.
THE IMPACT
Data science team achieved 65% faster development and resource efficiency
Because the data scientists have all of the analytics tools they need in a single solution, the team has accelerated development timeframes and reduced required resources by 65%—thanks to the modular, reusable development components, the centralized code repository, standardized tracking, monitoring, debugging, and deployment processes, and built-in audit reporting.
The new streamlined MLOps platform has also enabled the data science team to take a consultancy approach to analytics by creating a pipeline development roadmap that aligns with the investment firm’s strategic business goals.
A senior program manager at the investment company stated, “The impressive results that UST has achieved are game-changing. We have created a common investment language to inform the investment decision-making process. We are no longer debating the number or the method, but rather how we act on that number.”