Thesis Student - From Requirements to Proof: Exploring AI-Assisted Software Verification

From Requirements to Proof: Exploring AI-Assisted Software Verification- 30 Credits

Introduction

Interested in AI and software development? Explore how AI and logical reasoning could help check automotive software at TRATON. Apply your knowledge to a real engineering challenge and build connections for your future career.

Background 

Large language models (LLMs) can generate code from written requirements, but that code still needs independent validation. Specification-driven development (SSD) makes requirements precise so verification tools can check whether the code meets them.

At TRATON, Software System Change Requests (SSCRs) describe software changes. With many supported configurations, testing every combination can be costly and may miss edge cases. Formal verification uses logical reasoning to check code against specified requirements across the configurations included in a model. This could reduce the need for configuration-specific unit testing.

Objective

Explore whether formal verification, supported by an LLM, can show that a selected SSCR implementation meets its requirements across modeled configurations, without running tests as part of the check.

Job description

Starting with one selected SSCR and related User Function(s), you will explore how an LLM can help translate requirements into formal properties: precise statements that verification tools can check against the code. You will review generated specifications before verification. Compilation and static analysis may support the workflow. Existing test results can provide a comparison, but will not form part of the formal proof. You will assess what the approach can prove, its limitations, its suitability for the selected code, and SSD support for C++ and Java.

Your profile

You are studying towards a master’s degree in Computer Science, Software Engineering, Computer Engineering, Embedded Systems, or a related field. You enjoy programming and are curious about AI and software quality. Experience with C, C++, or Java is helpful.

Number of students:1 or 2

Start date for the thesis work: January 2027

Estimated time required: June 2027

Contact persons and supervisors 

Vaheed Nezhadali, Unit head, vaheed.nezhadali@scania.com

Marcus Jakobsson, Thesis supervisor, marcus.y.jakobsson@scania.com

Application: 

Your application must include a CV, personal letter and transcript of grades

A background check might be conducted for this position. We are conducting interviews continuously and may close the recruitment earlier than the date specified.