AWS ML Platform Starter Kit

Features

Infrastructure-as-code starter kit

Grid Dynamics has built and published a cloud-native starter kit on AWS Marketplace that accelerates the development and deployment of a modernized Machine Learning Platform. The starter kit enables rapid provisioning of the Machine Learning Platform, provides a CI/CD framework and sample models for reference.

Introduce experiments tracking

A managed Machine Learning Platform on AWS means you have a unified portal to manage experiments with the ability to track model-specific parameters, compare model versions and visualize model key attributes.

Production-ready Machine Learning Platform

Our Machine Learning Platform Starter Kit for AWS provides a full spectrum of ML capabilities: feature store and model repository, experiments tracking, serving layer for batch and real-time inference, model CI/CD.

The platform starter kit leverages Sagemaker for model development, Mlflow for experiments tracking, S3 as a feature store, EKS or Sagemaker for real-time serving and Amazon EMR (Elastic MapReduce) for batch serving. CI/CD automation is implemented on top of CodeBuild and Lambda, and platform deployment is fully automated on CloudFormation.

Use Cases

Build an ML Platform from scratch

Upgrade an existing ML Platform

Introduce new technical features

Migrate to the cloud

Reduce time-to-market

Adopt DataOps best practices

Industries

IoT & Manufacturing

The goal of using Machine Learning for smart manufacturing is to identify opportunities to improve industrial operations and OEE at any phase of the manufacturing process. A Machine Learning Platform empowers manufacturers to reduce costs, improve quality, detect anomalies, predict maintenance needs, accelerate factory floor processes, and achieve sustainability objectives.

E-commerce

Retailers in today’s digital age need to have their finger on the pulse of consumer behavior to stay ahead of the competition. Leveraging Machine Learning, retailers are more empowered to provide personalized customer experiences, optimize inventory and order management processes, predict demand for products, optimize pricing and marketing campaigns, and a number of other decision-support applications.

Ad tech

Technology and media companies can leverage Machine Learning to reveal diagnostic, predictive and prescriptive insights that help with content optimization, distribution strategies, marketing spend optimization, customer personalization, churn mitigation, and overall process efficiency by increasing the agility of data science teams.

Financial services

Machine Learning has become critical in the financial services industry for process automation, document analysis, portfolio management, algorithmic trading, risk management, fraud detection, money laundering prevention, customer experience personalization, pricing and product optimization, and more. In a nutshell, Machine Learning can analyze millions of data sets within a short time to improve financial services outcomes.

Healthcare

Machine Learning can save pharma and medicine hundreds of billions of dollars annually because of greater efficiencies in research and development and clinical trials, better insights for decision-making and more personalized customer/patient experiences, and innovative tools that can help consumers, physicians, regulators and insurers make informed decisions.

Benefits of an Analytics Platform in the cloud

Speed-to-market

With Infrastructure-as-code, you get to focus on building business critical applications, not on infrastructure provisioning and management.

Managed environment

Fully managed by AWS CloudFormation and AWS services, infrastructure provisioning takes days, not months.

Reduced operational costs

The Machine Learning Platform’s scalability and modular architecture enables flexible extension of new capabilities with serving on Amazon EKS (Elastic Kubernetes Service),or integration with 3rd party ML services.