Postdoctoral Position in Software Engineering: Program Analysis and Repository Mining

1000scholars; Odense; On-site; DKK 420,000 - 540,000; Full time

Job summary

SDU - University of Southern Denmark invites applications for a 2-year postdoctoral researcher position in the Software Engineering section of the Maersk Mc-Kinney Moller Institute (MMMI) in Odense, Denmark, starting February 1, 2027. This project investigates test-free program repair by using code-based evidence.

The successful candidate will design, implement, and evaluate a framework combining program analysis with mined software patterns across multiple languages, including AI-assisted code.

Qualifications

• PhD (completed or close to completion) in Software Engineering, Computer Science, Programming Languages, AI or closely related field.
• Strong interest or experience in program analysis, mining software repositories, automated software engineering, software testing or software repair.

Responsibilities

• Investigate whether software defects can be repaired without relying on test suites or specifications.
• Contribute to design, implementation, and evaluation of a framework combining program analysis with patterns mined from existing software.
• Perform experiments across multiple programming languages, including AI-assisted code, and assess real-world bugs and AI-modified code.

Skills

Program analysis

Mining software repositories

Automated software engineering

Software testing

Software repair

English communication

Independent collaboration

Education

PhD (completed or near completion) in Software Engineering, CS, Programming Languages, or AI

Tools

C

Java

Python

Job description

Postdoctoral Position in Software Engineering: Program Analysis and Repository Mining

SDU - University of Southern Denmark

Odense, Denmark

A 2-year postdoctoral researcher position is available at the University of Southern Denmark (SDU) in the Software Engineering section of the Maersk Mc-Kinney Moller Institute (MMMI). The successful candidate will contribute to the VILLUM Experiment project “Is Correctness Necessary? An Experiment in Test-Free Program Repair.” The project investigates whether software bugs can be repaired using information found in the code itself, without relying on test suites or specifications. The position is available from February 1, 2027.

Project Description

Generating code is becoming cheap; establishing whether it is correct is not. Large language models can now produce and modify software at remarkable speed, and practices such as “vibe coding” are widening the gap between code generation and validation. Requiring a carefully designed test suite or specification for every generated fragment is unlikely to scale. Yet most automated repair techniques depend on exactly such a correctness oracle: without a failing test or specification, they have neither a clear repair target nor a way to assess a proposed change.

This project challenges that dependency by investigating whether software bugs can be repaired using evidence already present in the code. It will combine program analysis with patterns mined from existing software to explore whether multiple, independent sources of evidence can guide repair when tests or specifications are unavailable. The approach will be studied across different programming languages and evaluated on both real-world bugs and code modified by AI coding assistants. The broader aim is to determine when code contains enough information to support its own repair and when knowledge of the intended behavior remains essential.

What we expect

The appointed researcher will investigate whether software defects can be repaired without relying on test suites or specifications. The candidate will contribute to the design, implementation, and evaluation of a framework that combines program analysis with patterns mined from existing software. The work will involve experiments with real-world bugs and code modified by AI coding assistants across multiple programming languages. We are looking for a highly motivated candidate who meets the following qualifications and characteristics:

• A PhD (completed or close to completion) in Software Engineering, Computer Science, Programming Languages, Artificial Intelligence, or a closely related field;


• Research experience or a strong interest in one or more of the following areas: program analysis, mining software repositories, automated software engineering, software testing, or software repair;


• An interest in software reliability challenges arising from AI-generated and AI-modified code;


• Experience with empirical evaluation, software repositories, or bug benchmarks is an advantage;


• Experience with one or more of C, Java, and Python is an advantage;


• Proficiency in written and spoken English, with good communication skills;


• Ability to work independently and collaboratively within a research team;


• A strong research record or demonstrated research potential through publications or other relevant research outputs.

Deadline 15 November