DeepCode’s AI Code Review: A next generation of static analysis tool with over 300K patterns
Earlier this year DeepCode AG publicly launched its AI Code Review service, with a free offer for Open Source repositories in GitHub. The service utilizes DeepCode’s innovative AI Software Platform that combines Big Code, cutting-edge program analysis methods and proprietary representation model enabling the AI Software Platform to use powerful machine learning methods over the code. The DeepCode’s serving infrastructure can offer a multitude of services that no existing tools, services, and solutions can match.
The AI Code Review service
When compared to the vast majority of the code review services, which are only based on simple and open source lint tools and a limited number of manually created rules, DeepCode can provide hundreds and thousands of unique suggestion through its language-independent AI Software Platform. This allows DeepCode to capture a much wider range of software defects and code style issues than any other code review system.
As opposed to lint tools, DeepCode avoids being annoying by providing only suggestions that other developers did to fix bugs and improve the quality of open source software. And the best part — when new learnings, developments or trends are implemented by the developer community, DeepCode’s AI Software Platform quickly learns those and suggests them to the AI Code Review users having similar code in their repositories. The AI Code Review service is saving the valuable developer time from manual code reviews as well as offering unique recommendations on performance, security, style and many other categories of suggestions.
The mission of DeepCode is to enable developers to work faster, deliver high-quality code and ultimately delight users with good and secure services. The next frontier is code synthesis, which we plan to initially offer in the form of automatic code fixing for some of the suggestions our AI Code Review service provides.
About DeepCode AG
The company was founded in Zurich, Switzerland as an ETH Zurich Spin-off and two of its founders (Prof. Dr. Martin Vechev and Dr. Veselin Raychev) are the widely published and awarded global thought-leaders in program analysis, algorithms and machine learning. Both of them are winners of multiple prestigious awards. The doctoral dissertation ”Learning from Large Codebases” by Dr. Raychev was honored by ACM as one of the best computer science dissertations in 2016.
In 2018 DeepCode received an overwhelming interest from investors on its seed financing round, ultimately btov Partners, one of the largest and oldest venture capital firms in Europe, joined forces with DeepCode to deliver the next step-function increase in the abstraction level in software development.
