Machine Learning with AWS Fargate


Hero Banner Description Text

AWS Fargate.png

Machine Learning and Artificial Intelligence with AWS Fargate

Machine Learning and Artificial Intelligence (ML/AI) are becoming important assets by automating applications and environments, and serverless machine learning is enabling organizations to become more efficient and to offer higher quality outputs than what has existed in the past. AWS Fargate automates the provisioning of the infrastructure for containers, creating efficiency in deploying containers. It works with services like Amazon SageMaker and AWS CodePipeline to enable companies to implment containerized machine learning applications. 

Read below to learn more about how ML/AI with AWS Fargate can be beneficial, and see how Veritone and Corteva Agriscience were able to leverage these capabilities.


fargate diagram.png



Machine Learning with Amazon SageMaker and AWS Fargate

Amazon SageMaker enables every developer to build, train, and deploy machine learning models quicky - it makes ML more accessible by taking care of the workflow. It will label and prepare the data, choose an algorithm, train the model, tune and optimize it for deployment, make predictions, and take action. 

Similarly, AWS Fargate takes care of the infrastructure of containers, leaving more capacity to work on the applications and tasks. Using both Amazon SageMaker and AWS Fargate, you can easily run containerized machine learning applications without thinking about the underlying infrastructure. Below is a demonstration of how these products can work together to build a visual search engine, and you can learn more about building this search engine on GitHub.



Customer Story: Veritone Builds Real-time Artificial Intelligence with AWS Fargate

Veritone's platform is designed to ingest data through a series of batch processes that are called "engines", and attach an output to it. For example, it can take in video and attach an output of trancripts or facial recognition. From here, the next goal was to design a data pipeline that would enable them to process the content in real-time with use cases from live TV to public safety. 

With AWS Fargate, Veritone was able to use Docker containers for their internal services and cognitive engines - it brought flexibility to their deployment. To learn more about the implementation, see the diagram below and read the full blog here.


Customer Story: Corteva Agriscience uses AWS Fargate and AWS CodePipeline for Machine Learning

As the agricultural division of DowDuPont, Corteva Agriscience supprts a network of research stations to improve agricultural productivity. A new way that they are working toward this increased productivity is by using machine learning to score genetic markers. Prior to implementation with AWS, technicians had to score the genotypic assays manually, but now have the ability to priortize other areas, such as technology development.

In order to bring the project into the production environment, Corteva Agriscience used the following AWS services:

  • AWS CodeCommit, CodePipeline, and CodeBuild are used for the CI/CD tooling.
  • CloudFormation is our preferred method to describe, create, and manage AWS resources.
  • Amazon Elastic Container Registry (ECR) is used to store the needed Docker container image.
  • AWS Systems Manager Parameter Store is used to hold secrets like database passwords.
  • AWS Fargate is used for the actual application stack, obviously.

Learn more about the implementation with Corteva Agriscience here



Learn more about AWS Fargate, Amazon ECS, and Containers on AWS!

Security on AWS Fargate

Observability on AWS Fargate

Serverless Architecture with AWS Fargate

CI/CD with AWS Fargate