Master Thesis: Data Structure on LLM-Based Verification Analysis

Join our Team

About this opportunity:

Large Language Models (LLMs) and AI agents are creating new opportunities for engineering and hardware verification. A key challenge is determining whether verification data should be provided as raw, unstructured logs or transformed into structured representations to enable reliable, accurate, and efficient analysis.

In this thesis, you will experimentally compare how different levels of data structure affect LLM-based verification analysis:

Raw verification logs → Partly structured data → Fully structured data

You will work with real-world verification data and evaluate how these representations influence analysis quality and failure diagnosis under comparable conditions.

The thesis should provide evidence-based recommendations on how verification data should be represented and structured to support reliable LLM-based analysis in hardware verification.

What you will do:

• Work with real-world hardware verification data and test logs
• Define and create raw, partly structured, and fully structured representations of verification information
• Develop consistent ways for LLMs to access each data representation
• Design representative analysis and failure-diagnosis tasks
• Perform controlled and reproducible experiments
• Evaluate accuracy, reliability, context usage, latency, token usage, cost, coverage, and potential information loss
• Provide recommendations on when and how verification data should be structured for AI-assisted analysis

Potential extension topics may include comparing multiple LLMs, evaluating robustness against incomplete or noisy verification data, identifying which aspects of structuring create the greatest value, and comparing structured-data approaches with retrieval-based solutions such as semantic search, databases, and MCP-based integrations.

The focus is on experimental evaluation and research, rather than building a complete production AI platform.

The thesis should provide evidence-based recommendations on how verification data should be represented and structured to support reliable LLM-based analysis in hardware verification.

The goal is not simply to demonstrate that an LLM can analyze test data, but to determine when structured data outperforms raw logs, which level of structure is most useful, and what trade-offs the transformation introduces.

The skills you bring:

• Interested in combining AI research with real engineering challenges
• Large Language Models and Agentic AI
• Python and data analysis
• SQL/PostgreSQL and databases
• Software or hardware verification
• Test automation
• Experimental evaluation of AI systems

Why join Ericsson?At Ericsson, you'll have an outstanding opportunity. The chance to use your skills and imagination to push the boundaries of what's possible. To build solutions never seen before to some of the world's toughest problems. You'll be challenged, but you won't be alone. You'll be joining a team of diverse innovators, all driven to go beyond the status quo to craft what comes next.

What happens once you apply? Click Here to find all you need to know about what our typical hiring process looks like.Encouraging a diverse and inclusive organization is core to our values at Ericsson, that's why we champion it in everything we do. We truly believe that by collaborating with people with different experiences we drive innovation, which is essential for our future growth. We encourage people from all backgrounds to apply and realize their full potential as part of our Ericsson team. Ericsson is proud to be an Equal Opportunity Employer. learn more.

Primary country and city: Sweden (SE) || Stockholm

Req ID: 791322