Leveraging machine learning for assisting game developers on their software telemetry
Keywords that describe the project: DevOps, Game development, AI
This project is the intersection between three exciting and emerging topics, which are related to applying artificial intelligence (machine learning in particular) methods to the DevOps practices in the specific context of games development. In particular, this project focuses on applying machine learning techniques to help game developers improve the traceability (a fundamental part of DevOps) of their games. The student will develop a software system that leverages machine learning to recommend to game developers how they should add telemetry to their projects in a way that they trace enough information without sacrificing the game performances.
Precisely, the student will mine the repositories of three game projects to understand how they collect telemetry, collect metrics to train machine learning models for telemetry recommendation, and evaluate these models.
We are looking for students with good programming skills (python, R, Java, Git), and knowledge on machine learning (e.g., logistic regression, random forest, …).