Based in 2016 as an ETH Zurich spinoff DeepCode developed a platform for automated AI code reviews which alerts developers about critical vulnerabilities in software projects, hence helping them to prevent bugs from going into production. Developers get started by connecting DeepCode with their GitHub, BitBucket, or GitLab accounts or directly within their integrated development environment (IDE), therefore no configuration is required.
DeepCode brings knowledge from the global development community to every software developer. The bot uses machine learning to semantically analyze hundreds of millions of commits in numerous open-source software projects and learns how to find millions of serious coding issues. It identifies bugs before they happen so that developers can fulfil their true coding potential and deliver quality products faster.
To date, more than four million developers worldwide use the bot to find bugs and security vulnerabilities in software projects. The developers contribute to the software repositories on which DeepCode based.
“Our data shows that over 50% of repositories have critical issues and every second pull-request has warnings about issues that need to be fixed,” said Boris Paskalev, CEO and co-founder of DeepCode. “By using DeepCode, these issues are automatically identified and logically explained, and suggestions are made about how to fix them before code is deployed.”
Achievements and prospects
As part of their achievements, DeepCode found two major bugs in the Swiss COVID-19 contact tracing app, and for the second year in a row, it received the European Commission’s Seal of Excellence for Research and Innovation.
DeepCode is now shifting development towards Augmented Intelligence by using the knowledge of the global developer community to enhance the experience of every single developer and help them learn and deliver better code faster, as revealed by the startup.
“Our next milestone is to dominate the security space and to support major languages and integrations that users need. Some new services are in progress based on the unique knowledge our platform holds and its ability to semantically represent in real-time any code construct during the whole development history of a given language—from the very first line of code ever written”, said Paskalev.