Github Reviewer Recommendation for incoming Pull-Requests(PR) using Graph Neural Network

Github Reviewer Recommendation is a tool for the larger projects on the platform like GitHub, Collaboration is done by cloning the repository, modifying the sources, and then issuing the pull-requests. For big scale projects, the assignment of reviewers becomes a challenging task. For example, multiple reviewers might review the same pull-requests, a reviewer may check the pull-request from out of his/her domain. That may lead to a considerable feedback loop time, and sometimes the critical changes might be ignored. So, we are building a recommendation system for recommending reviewers to the pull requests so that such a system can save reviewers’ time and improve overall code quality due to a short feedback loop time. In this work we are trying to build the GitHub Reviewer Recommendation System for the Kubernetes repository on GitHub and trying to generalize it for other large projects. We will be building a Graph Neural Network and compare its performance with existing Supervised Random Walk implementation.