Tolga Talks Tech is a weekly summer video series in which Onica’s CTO Tolga Tarhan tackles technical topics related to AWS and cloud computing. This week Tolga talks about Amazon SageMaker with Onica’s VP of Solutions and Development, Eric Miller. For more videos in this series, click here.
What is AWS SageMaker?
Amazon has been doing some interesting things in the Artificial Intelligence(AI) and Machine Learning (ML) space, the most important of which is their newest flagship ML tool: Amazon SageMaker. Amazon SageMaker is an end-to-end platform for designing, developing, training and deploying machine learning models on AWS. It’s a fully managed AWS service that enables you to quickly and easily integrate machine learning-based models into your applications.
How do you use Amazon SageMaker?
You can use AWS SageMakers to build, train, and deploy Machine Learning models. Here’s how to make it happen:
- Create an AWS Account
- Create an Amazon S3 Bucket
- Create an Amazon SageMaker Notebook Instance
- Create a Jupyter Notebook
- Download, Explore, and Transform Data.
- Train a Model
- Deploy the Model
- Validate the Model
- Clean Up
For more details on using SageMaker, review the Amazon documentation.
How is AWS SageMaker different than what’s been done before in the AI/ML space?
Amazon SageMaker is distinct in that it lowers the bar of accessibility for people who are not data scientists but really want to build useful ML functionality into their platforms. Amazon SageMaker allows us to make things easier without losing the functionality and flexibility of real Data Science tooling.
With Amazon SageMaker, you still get exposed to real data science tools like deep neural networks, linear learners, k-means algorithms, and still have access to tuning hyperparameters. AWS SageMaker can be summarized as “DevOps for data science, as it follows many of the same Deployment Pipeline patterns we see used by development teams when delivering software solutions. This allows us to use a Docker image to create a local development environment, develop, train and tweak our model locally, then push that model to SageMaker for CPU intensive training with GPU instances and clusters. If we like what that model is doing then we can deploy it out through dev, test and production environments just like we do with a software solution
Data is the new currency. Onica can help you capture this by enabling your Data Science team with AI/ML solutions powered by AWS products like Sagemaker, allowing your team to focus on building models rather than waiting for them to be solved. Contact us to learn more!