I'm starting my fourth year of Computer Science and need to choose an undergraduate thesis topic related to DevOps or cloud computing. I'd like to carry out the project in collaboration with the company where I'm working, but they asked me to propose a few possible ideas first. I'm still fairly new to DevOps and cloud—I started my internship in July—so I'd prefer something manageable rather than highly complex. The company uses Azure, Terraform, GitHub Actions, Docker, Linux, Python, Bash, and PowerShell. I've also worked a little with Azure monitoring, Azure Data Explorer, cloud networking, and security. Kubernetes is used internally as well, and an upcoming project involves fleet data. I'm mainly hoping to work with real company processes, gain practical experience, and produce a thesis with measurable results. What topics might fit this situation?
4 Answers
Other reasonable directions would be improving CI/CD reliability, analyzing deployment failures, or evaluating monitoring for cloud workloads. Kubernetes could be worth exploring if the company already has a suitable use case, but I’d avoid making the thesis about introducing a large new platform unless there is a clearly defined problem to solve. The best topic will probably come from asking a team which recurring task costs them the most time or money.
Security in the CI/CD pipeline could also make a manageable thesis. You could integrate tools such as Trivy for container scanning and Gitleaks for detecting accidentally committed secrets, then measure how many issues are caught and how much extra build time the checks add. That gives you a clear prototype, evaluation criteria, and practical value.
A good way to narrow this down is to focus on a measurable improvement for the company rather than choosing a technology first. For example, you could study cloud cost optimization by automatically shutting down non-production environments outside working hours and cleaning up unused resources. You could measure the actual monthly savings and evaluate whether the automation causes any operational problems.
Standardizing local development environments with Dev Containers might be another approachable option. Compare the existing setup process with a containerized setup by measuring onboarding time, configuration problems, and consistency between developers. It is easier to scope than a large platform project while still being relevant to the company’s development workflow.
These are really actionable suggestions. I especially like that each one has a concrete metric I could evaluate instead of being just a general technology investigation.