Comprehensive software reviews to make better IT decisions

Amazon Releases SageMaker Studio – An IDE for Machine Learning

Last month, Amazon released SageMaker Studio, an interactive development environment (IDE) for machine learning (ML). The objective for this new-ish offering was to address immature tooling in ML and make it easier for data scientists to create and deploy ML models.

SageMaker Studio is an expansion of Amazon’s SageMaker, a fully managed ML service introduced two years ago to enable developers to “quickly build, train, and host machine learning models at scale.” SageMaker is used by ADP, Dow Jones, Intuit, GE Healthcare, Celgene, Zalando, Tinder, and Hotels.com, among many others.

Source: Amazon SageMaker Studio: The First Fully Integrated Development Environment For Machine Learning, Amazon Blog. Published December 3, 2019.

IDEs are software applications that are widely used in traditonal software development. They integrate a variety of tools – source code editors, debuggers, compilers, version control, collaboration, and monitoring – which make developers’ work easier, increasing their efficiency and productivity, enhancing collaboration, and ultimately resulting in better-quality applications.

A lot of such tools have been lacking for ML, or – when available as open-source tools or a vendor offering – they have to be patched together. “Amazon SageMaker Studio unifies at last all the tools needed for ML development. Developers can write code, track experiments, visualize data, and perform debugging and monitoring all within a single, integrated visual interface, which significantly boosts developer productivity,” writes the announcement blog. Specifically, SageMaker Studio includes:

• Amazon SageMaker Notebooks to create and share Jupiter notebooks.
• Amazon SageMaker Experiments to organize, track, and compare thousands of ML jobs.
• Amazon SageMaker Debugger to debug and analyze complex training issues and to receive alerts.
• Amazon SageMaker Model Monitor to detect and visualize quality issues such as data drift and to receive alerts.
• Amazon SageMaker Autopilot to build models automatically, including algorithm selection and data preprocessing, and provision the required infrastructure.

The Amazon blog announcing the IDE contains several screenshots of these utilities and more details on each of them.

Our Take

The reaction to SageMaker Studio has been mixed, from headlines claiming that the IDE is not yet ready to win over data scientists to positive reactions. While some of the criticism is perhaps warranted – from product complexity to repackaging services and utilities that had been offered for some time as new – unifying these services into a single environment is a welcome development that will be met with enthusiasm by many data scientists.

Want to Know More?

Read reviews and contribute your own to our new SoftwareReviews category Machine Learning Platforms.