Many CyPerf QA teams still rely on legacy Jenkins jobs defined in XML format. While functional, these jobs are hard to review and don’t align with modern CI/CD best practices, which recommend storing pipelines as code using Jenkinsfiles. Migrating these jobs manually is slow, error-prone and requires significant effort.
Automating this migration using AI would significantly reduce manual effort, improve pipeline consistency and accelerate CI/CD modernization across teams.
This requirement is especially important for our planned migration to CloudBees, a Jenkins fork that meets the latest security requirements and is managed by a third party rather than in-house. Moving to CloudBees would reduce our operational burden, but migrating legacy jobs is a major blocker, one that this solution would directly address.
What You’ll Work On You’ll help automate the Jenkins job migration process using AI (LLMs):
- Use AI/LLMs to convert legacy Jenkins job XML configurations into modern Jenkinsfiles.
- Analyze job definitions including build steps, triggers, parameters, credentials, and post-build actions.
- Generate clean, standardized, and best-practice-compliant Jenkinsfiles.
- Produce validation output and reports highlighting potential issues or ambiguous conversions.
- Support batch processing of large numbers of jobs and provide a CLI (with future UI extension).
Since real-world Jenkins configurations can be complex and vary widely, feeding raw XML directly to an LLM is often insufficient. We expect a structured, agent-like system that can intelligently parse, normalize, and contextualize job definitions before generating accurate pipeline code.
What you will gain:
- Exposure to DevOps tooling and CI/CD pipelines
- Hands-on experience with AI-assisted code generation
- Understanding legacy system migration and automation best practices
- Opportunity to learn prompt engineering and AI workflow integration.
Skills required: Python programming, XML parsing and code analysis, DevOps concepts and Jenkins pipelines, familiarity with AI / LLM integration (OpenAI API or local models), basic CLI / basic web development for UI
