Thesis: Don't Trust the Logs: Containing AI Agents in Azure Pipelines (AI)

Location: Onsite at the hirify.global office in Sundsvall, Sweden. Start date: January or March 2027.

Company

hirify.global is a strategic partner in product and digital service development, combining cross-functional expertise with business understanding to deliver solutions from idea to completion.

What you will do

• Build a simple failure-triage AI agent that diagnoses failed CI/CD builds from unstructured logs.
• Run the unchanged agent in the pipeline agent environment, Azure Container Apps, Container Apps Sandboxes, and optionally Foundry Hosted Agent.
• Compare startup latency, execution duration, cost per triaged failure, reliability, and diagnostic quality using repeated runs.
• Test prompt-injection scenarios involving attempts to leak environment variables, call external URLs, or access files outside the working directory.
• Measure root-cause accuracy against a non-AI baseline using ground truth from two independent labelers.
• Keep the setup reproducible with Bicep, a pinned model version, and recorded prompts.

Requirements

• Bachelor’s or Master’s student in Computer Science, Computer Engineering, Data Science, or a related field.
• Graduation in June 2027.
• Fluency in Swedish, both written and spoken.
• Fluency in English, both written and spoken.
• Comfort with Git, basic CI/CD concepts, Python or a similar scripting language, and containers.
• Prior Azure experience is not required.

Nice to have

• No additional optional qualifications are specified.

Culture & Benefits

• Hands-on thesis work on a real project in a supportive and innovative environment.
• Azure budget and model API access are provided by the supervisor.
• No company data is used; testing takes place with failed open-source CI runs in a dedicated test subscription.
• The project excludes model training, production-grade deployment, and use of real company data.
• This is a thesis internship located at the Sundsvall office.

Hiring process

• Submit the application through the careers page only; applications sent by email are not accepted due to GDPR.
• Apply as soon as possible and no later than 30 November 2026.