ML platform with Sagemaker

LSports is a world-leading sports data company that provides an innovative sports betting data API for the sports betting industry. They are a leading provider of high-quality live sports data feeds, serving clients worldwide.

The Challenge

Lsports had a large data repository in their data warehouse and wanted to create machine-learning (ML) models from this data. They were looking for a platform that could pre-process, train, and deploy machine learning models quickly, and scalably, and streamline their ML development process.

The Solution

CloudZone had developed a full machine learning pipeline which was designed and built on top of the AWS SageMaker platform. The solution delivered a training pipeline that controls data preparation, model training, and model deployment. All stages were implemented on AWS SageMaker Pipeline, and the pipeline was wrapped in AWS native tools to control triggers and manual approvals. The data science team also integrated SageMaker Studio as an IDE to leverage purpose-built tools for ML development, like managing experiments, explainability capabilities, data visualization, and more. The goal of this solution was to streamline to MLOps in order to scale the solution and save costs.
This included design, development, testing, deployment to production, training and post launch support.

The Results

The customer has now gone through an enablement process that lets them step up their machine learning (ML) lifecycle and move to a higher level of automation with added capabilities for rapid innovation through robust machine learning lifecycle management. Some of the automations include:

  • Create reproducible workflow and models.
  • Easy deployment of high-precision models in any location.
  • Effective management of the entire machine learning lifecycle.
  • Machine learning resource management system and control.

The implementation of these advancements has yielded significant accomplishments in enhancing the customer’s machine learning initiatives:

  1. Improved efficiency: The project has led to a significant reduction in model training time, indicating that the new solution has made the machine learning process more efficient.
  2. Streamlined ML development process: The solution has also reduced the model deployment time, demonstrating that the new pipeline has streamlined the ML development process and made it faster.
  3. Enhanced scalability: The increase in the number of models trained and deployed after implementing the solution showcases the scalability of the new pipeline, allowing for more models to be developed and utilized.
  4. Better model performance: Working with a managed platform allowed us to create an environment that facilitates the development of improved models. This is achieved through the ability to test and experiment, which results in better models in terms of quality.

As a result of these advancements, Lsports has realized significant improvements, including:

  1. A remarkable 75% reduction in model deployment time.
  2. An impressive 50% reduction in the model development cycle time.

Daniel Netzer, Senior SA @Lsports:

“The Data team at CloudZone has been a pleasure to work with. They are quick and responsive, and always follow through on commitments. The development process for our machine learning pipeline was seamless and now we have a working ML platform with Sagemaker. It’s amazing how much we’ve progressed in just a few weeks!”