Amazon (AWS) SageMaker

Edited November 11, 2020 by Tarun Rawat and Clayton Winders and Andra Ferrara

Amazon SageMaker is a fully managed service that provides every developer and data scientist with the ability to build, train, and deploy machine learning (ML) models quickly

Amazon (AWS) SageMaker

What Is SageMaker?

Amazon SageMaker is a fully managed service that provides every developer and data scientist with the ability to build, train, and deploy machine learning (ML) models quickly. SageMaker removes the heavy lifting from each step of the machine learning process to make it easier to develop high quality models.

Who Should Use SageMaker?

  • When you need to perform machine learning using AWS cloud infrastructure.
  • When the you easy integration with other AWS services like S3.
  • When infrastructure compute or scaling is a roadblock for progress.

How Do I Learn About It?

Useful links to learn basics of SageMaker

  • https://aws.amazon.com/sagemaker/
  • https://aws.amazon.com/getting-started/hands-on/build-train-deploy-machinelearning-model-sagemaker/
  • https://docs.aws.amazon.com/sagemaker/latest/dg/gs.html
  • https://towardsdatascience.com/building-fully-custom-machine-learning-models-on-awssagemaker-a-practical-guide-c30df3895ef7

Strengths

  • Fully integrated development environment (IDE) for machine learning.
  • Studio Notebooks – one click Jupyter notebooks that can be spun up on AWS cloud with capability to scale the resources and compute power up and down without interrupting work.
  • Allow to automatically build, train and tune models
  • Easy integration with Amazon Ground Truth which another service that allows to build and manage highly accurate training datasets quickly.
  • Support available for leading deep learning frameworks like TensorFlow, PyTorch, Keras etc.

Limitations

  • It’s very good for the hardcore programmer, but a little bit complex for a data scientist or new hire who does not have a strong programming background.
  • Most of the popular library and ML frameworks are there, but we still have to depend on them for new releases.