A year ago, I used to find myself stuck with models that were only accessible in my Jupyter notebooks or localized API endpoints. While it was great for experimentation, it was difficult to scale my models and share them with others. That is until I discovered the power of cloud tools, and in this article we will talk about AWS Machine Learning (ML) tools.
In this 3–4 part article series, I want to share the importance of using cloud tools for Machine Learning Engineering (MLE). AWS ML tools provide scalable, efficient, and cost-effective solutions for data preparation, model training, and deployment. Whether you’re a seasoned MLE practitioner or just starting out, AWS ML tools can help you take your models to the next level. Let’s explore the possibilities and see how they can transform the way you work with ML!
Notes: This tutorial would assume that you already created an AWS account and the region selected (check the top right corner in your AWS Management Console) is US West (Oregon) us-west-2.
Before we start working with AWS ML, it’s important to set up an IAM user with the least privileged permissions necessary to complete the tasks in this article. IAM (Identity and Access Management) allows you to create and manage users and their permissions to access AWS services. There are a lot of tools and services within AWS and it is important for a user to only access the necessary tools for a project. By setting up an IAM user with least privilege, you can help ensure the security of your AWS resources.
- Sign in to the AWS Management Console and open the IAM console
2. In the navigation pane, choose Users, then click Add users
3. Enter a user name for the new user, then select the Programmatic access checkbox to give the user access to the AWS CLI. Keep everything else as default then click Next
4. In the permissions section, click Attach policies directly
5. In the permission policies, search for “AmazonS3FullAccess”, “AWSCloud9Administrator”, “AWSCloudShellFullAccess”, and “AmazonSageMakerFullAccess”. This gives the user to access Amazon S3, Cloud9, Cloud Shell, and SageMaker, click Next
Note: This is just an introduction on how to navigate the IAM service of AWS. This is not a secure way of creating an IAM user for a project or an organization. For more reference check this documentation of AWS for IAM user creation best practices including concepts such as least-privileged permissions.
6. In the Review and Create section, click on Create User
7. Securely save the user name and password of the created user and also take note off the console sign-in URL
Congrats! You successfully created an IAM user and now you can log in using those credentials into the console sign-in URL. Make sure to keep your temporary password before exiting the page.
Stay tuned for the next installment of our Getting Started with Machine Learning in AWS series, where we will dive into using S3 and Cloud9 for data preparation and model training. In Part 3, we’ll explore the power of SageMaker for model deployment and management. With these tools at your disposal, you’ll be able to take your machine learning projects to new heights. Don’t miss out!