Oleh Hlovatskyi

Sebastian Ewak
Article content
Amazon Codeguru for AI-Powered Code Analysis
An artificial intelligence-based code quality analyzer CodeGuru from Amazon Web Services (AWS) is now publicly available. This was announced by the company’s press service in early July 2020, highlighting that now it is more convenient and functional
TECHNOLOGY FEATURES
CodeGuru is a tools’ set that uses a heuristic neural network algorithm to view code. This software distinguishes over 8000 types of errors and offers real-time optimization to improve performance.
The set consists of an analyzer and a recommendation database. It was first launched as a limited access set of tools in January this year. The AWS Trained Reviewer (AI) uses data from over 12,000 open source projects as well as about 5,000 different types of Amazon native scripts.
“Amazon may be the largest organization, but it is also difficult to have enough experienced developers with enough free time to review the code. The amount of code that is written every day is huge.” – says an Amazon spokesman.
“And even the most skilled reviewers miss issues before they impact applications, leading to bugs and performance issues.”
When working with CodeGuru, developers are not tied to using Amazon’s CodeCommit repository service and can use any alternative they want, including GitHub and Bitbucket Cloud.
After making all adjustments, they can continue to commit their code to the chosen repository and CodeGuru will analyze it on an ongoing basis.
All suggestions are made in the context of the repository. The analyzer generates a pull request automatically adds a comment with additional information on errors or code improvements.
