Sagemaker vs. Data Science platforms

Question

What is the distinction (and overlap) in the positioning of Sagemaker vs. Data Science platforms such as KNIME or RapidMiner? I have a client who is asking for my opinion and can't find any documentation that clearly delineates between the Data Science platforms and ML aaS such as SageMaker.

AWS-User-9919757 · Accepted Answer

I have looked at KMine a bit the past couple of weeks and have used SageMaker in depth for a couple of POCs.

I like to frame this question in the context of Andy Jassy's three tiers of the AWS ML stack.  My view is that KMine fits into the top layer.  It's a drag'n'drop interface, suitable for those who want a push-button solution to an ML problem.

SageMaker fits into the middle layer.  It automates most of the heavy lifting around infrastructure, model training, and model deployment.  But you have to understand which algorithm to use and how to tune it.  It's more flexible than the push-button approaches but far easier than starting with a deep learning AMI.

Sagemaker vs. Data Science platforms

Relevant content