TrustRadius
Amazon SageMaker is quite good at machine learning, as long as you're in the Amazon Ecosystem.
https://www.trustradius.com/machine-learningAmazon SageMakerUnspecified7.813101
Thomas Young profile photo
February 04, 2019

Amazon SageMaker is quite good at machine learning, as long as you're in the Amazon Ecosystem.

Score 8 out of 101
Vetted Review
Verified User
Review Source

Overall Satisfaction with Amazon SageMaker

Amazon SageMaker is used by a specific department that supports machine learning models development and deployment. From my perspective, the software makes a valiant effort at making data mining and machine learning more user-friendly, something that is not always an easy job. SageMaker addresses clients who wish to use machine learning for market predictions, looking for data mining details, and predictive analytics. It's great for what it attempts to do well.
  • Amazon SageMaker is great for visually seeing the development of machine learning models. The process is set up in a logical step-by-step process.
  • Amazon SageMaker makes training data models about as easy as it gets. It's straight-forward to construct training and test samples.
  • Amazon SageMaker makes deploying machine learning models much easier than other open-source tools.
  • Amazon SageMaker is a great tool for a data scientist, although surprisingly, comparing different machine learning models with SageMaker is not as easy as one would think. I think Amazon needs to team up with a data scientist who does ensemble modeling.
  • Because SageMaker is targeted for machine learning models, other models a data scientist might use require more effort to get them incorporated. My guess is Amazon is moving to make SageMaker a more complete tool.
  • SageMaker can take a long time to run on larger data sets. That's the case with every big data science tool I've used, but SageMaker doesn't seem to be as quick as other tools.
  • SageMaker was a positive return on investment for those analysts that wanted to use new tools and had the aptitude to implement machine learning models with new resources. Such individuals were quick to improve their productivity with SageMaker.
  • SageMaker was not a useful tool for analysts used to running reports on small data. Machine learning is a new area, and analysts that are only marginally interested in machine learning will find SageMaker not worth the effort.
  • When used in combination with other machine learning tools, SageMaker is a perfect addition because it complements the Amazon ecosystem while the other tools provide more simple model development environments.
Amazon SageMaker is the best option for machine learning if you are already using the Amazon data science ecosystem. The software integrates nicely with MapReduce and most of the other Amazon tools. Additionally, MapReduce does a fairly good job of making the development of machine learning models a user-friendly experience. Sometimes that is no easy task. With this said, it is clear when using MapReduce that the Amazon developers do not have as long of a history in advanced analytics, something that competing tools, such as IBM Watson, has a complete grasp on. Other tools are similar in speed and functionality.
IBM Watson Analytics, IBM Analytics Engine, IBM Watson Campaign Automation, IBM Watson Content Analytics, Watson Studio (formerly IBM Data Science Experience), Google Cloud AI, Amazon Tensor Flow, TensorFlow, Google Ad Manager, Google Analytics Premium, Google Cloud Dataflow, 6sense, MapR, Cloudera Data Science Workbench, Cloudera Enterprise
Amazon SageMaker is a great tool for developing machine learning models that take more effort than just point-and-click type of analyses. The software works well with the other tools in the Amazon ecosystem, so if you use Amazon Web Services or are thinking about it, SageMaker would be a great addition. SageMaker is great for consumer insights, predictive analytics, and looking for gems of insight in the massive amounts of data we create. SageMaker is less suitable for analysts who do generally "small" data analyses, and "small" data analyses in today's world can be billions of records.