Master Thesis: AI-generated Golden Paths for Platform Engineering

Long-lived credentials in R&D production systems, CI/CD pipelines and service integrations can create security risks through exposure, misuse or excessive validity. This thesis at Ericsson investigates how workload identity federation, Just-in-Time (JIT) and Just-Enough Access (JEA) can reduce persistent credentials and standing privileges. It also explores AI-assisted risk assessment for continuously adapting workload access.

The assignment

• Investigate and evaluate an architecture connecting workload identity, federation, JIT/JEA authorization, continuous risk assessment and adaptive enforcement.
• Determine which federation patterns are applicable to R&D production environments and how JIT/JEA can be used for non-human identities.
• Identify signals that should prompt a reassessment of workload trust.
• Examine whether AI-assisted analysis can support automatic access restrictions or revocation while preserving explainability, predictability and reliability.
• Develop a reference architecture and governance model for continuous workload trust, backed by empirical evaluation of security and operational characteristics.

Who we are looking for

• Currently studying for a master's degree in Computer Science, Information Security, Software Engineering, Data Science or a related discipline.
• Strong programming ability in Python, Java or a comparable language.
• Interest in cybersecurity, secure software engineering, AI/LLMs and workflow automation.
• Experience with Git/GitLab, CI/CD, security assessments or n8n is beneficial.
• Able to work independently and methodically, with clear communication in English.

What we offer

• An opportunity to apply technical skills and creative thinking to challenging problems alongside a team of innovators.
• A collaborative environment with support from colleagues.

Practical details

Location: Linköping. Suitable for one or two students, with a preference for two. Application reference: 792245.