SECS Micro-Intern: Academic Research | Oakland University

The role will include a research project in the Computer Science and Engineering Department.

In this project, you will explore whether a pre-trained AI language model — the same type of technology powering tools like GitHub Copilot — can automatically detect security vulnerabilities in real-world source code.

Working with publicly available vulnerability datasets, you will fine-tune a CodeBERT model using Python and the HuggingFace library, evaluate its performance, and write up your findings in a short report.

No prior experience in AI or cybersecurity is required; basic Python knowledge and curiosity are all you need.

The project is designed to be completed in approximately 40 hours and will give you hands-on experience with modern AI tools, real security datasets, and the kind of research workflow used in top academic labs and industry settings.