We’ve been working on using machine learning for software engineering since 2010. An overview of the research questions in this area is presented in the figure below.
Decision trees to speed up software development
Using Machine Learning for defect predictions
A big chunk of our work is about predicting defects. We can predict the number of defects per week, their location, severing and which team should be assigned to fixed it. We even have gadgets and apps to support quick information sharing.
Using machine learning to automate code reviews
In 2016, we created a completely new way of finding violations in coding guidelines. We developed and evaluated a tool for that – CCFlex.
The tool is available on github: https://github.com/mochodek/py-ccflex