Doctoral student in Software Engineering Full-time

Doctoral Student In Software Engineering

Third-cycle subject: Computer Science

The doctoral work will be done as part of research project “AI for Code”.

This PhD project develops a mechanistic interpretability framework to assess and improve the correctness of code generated by large language models (LLMs). Rather than relying on post-hoc behavioral validation through test suites, the project identifies internal “correctness signals” within LLM activations using Sparse Autoencoders (SAEs) for feature decomposition, causal inference for validation, and activation steering to actively guide code generation. The work is organized around three research objectives: decomposing superposed neural representations into interpretable features, modeling the causal relationships between those features and code correctness outcomes, and applying targeted steering interventions to reduce code failures. The project is grounded in prior work on neural program repair and execution-based learning signals, and it will be evaluated on standard benchmarks including HumanEval and BigCodeBench. The expected outcome is an end-to-end interpretability pipeline that provides human-understandable explanations for why an LLM succeeds or fails at generating correct code, with direct relevance to the safe deployment of AI coding tools in high-stakes domains such as healthcare, finance, and defense.

The candidate should have a strong background in computer science. The PhD student is expected to publish in the top conferences and journals in the field of Software Engineering. The working language is English.

Supervision: Martin Monperrus is proposed to supervise the doctoral student. Decisions are made on admission.

To be admitted to postgraduate education (Chapter 7, 39 § Swedish Higher Education Ordinance), the applicant must have basic eligibility in accordance with either of the following:

• passed a second cycle degree (for example a master's degree), or
• completed course requirements of at least 240 higher education credits, of which at least 60 second-cycle higher education credits, or
• acquired, in some other way within or outside the country, substantially equivalent knowledge

In addition to the above, there is also a mandatory requirement for English equivalent to English B/6.

In order to succeed as a doctoral student at KTH you need to be goal oriented and persevering in your work. During the selection process, candidates will be assessed upon their ability to:

• independently pursue his or her work
• collaborate with others,
• have a professional approach and
• analyze and work with complex issues.

After the qualification requirements, great emphasis will be placed on personal skills.