Every developer knows the pain of code reviews – a huge bottleneck for tech teams. Developers spend 20% of their time reviewing their team’s code before it is added to the codebase, leading to days of back and forths.
Meet Quack AI. Using deep learning, we automate up to 80% of the code review process, so that developers can focus only on the parts that need a human touch. To do this, we’re pioneering proprietary modelling approaches to code. They will increase computational efficiency beyond the current state of the art, while avoiding to neglect the complexity of human interaction and intent involved in code.
We currently have 14 partners who’ve given us access to their review data, so that can train and build their custom code reviewers. We’re accompanied by 2 high-profile tech advisors (including a leading researcher at Deep Mind), to build our pioneering R&D approach to code.
