JOB DESCRIPTION

You will research the state-of-the-art in AI and apply it to real-world software problems at our industrial partners: ASML and DCODIS (a start-up). This is technically challenging applied research with as main outcome a proof-of-concept tool that allows developers to quickly find and fix software errors including security vulnerabilities. You will innovate the Find2Fix pipeline by making the different steps, including found issues and suggested patches, easier to understand using interpretable AI using state machine models and LLM-based explanations. You will provide the community with the first tool for self-healing software that is useful for research, education, and industrial use. Your research will be published and presented at international AI, software engineering, and security venues.
The first position is focused on discovering new errors using behavioural models learned from software. You will extend fuzzing technology by learning such models in real-time and using them to: guide the fuzzer, discover logical faults, and pin-point their root cause. This builds on the works of the cyber analytics lab (www.cyber-analytics.nl) and CISElab (www.ciselab.nl) at TU Delft.
The second position is focused on finding patches for already discovered errors. You will research and extend template- and LLM-based program repair using modern program synthesis techniques based on the Herb toolkit (herb-ai.github.io).
The project is a close collaboration between the labs at TU Delft and the digital security and software science departments at Radboud University, where two additional PhD students will work on the Find2Fix project. Together with the rest of the team, you will build demonstators for the Find2Fix technology at our industrial partners.
You will work in the cyber analytics and CISE labs in the Algorithmics and Software Engineering Research groups at the Software Technology department under supervision of dr. Sicco Verwer, dr Annibale Panichella, and dr. Sebastijan Dumancic. The Algorithmics group aims to design, and understand fundamental properties of algorithms for intelligent decision making in real world applications, such as for coordinating electrical loads within network constraints, planning logistic processes on a container terminal, or finding the root-cause of a system failure. Apart from dealing with the scalability challenge, we work on two main research objectives. First, we aim to understand how to learn models that are useful for upstream planning, scheduling, and verification tasks, and how to deal with their uncertainty. Our second research objective is to understand how to effectively incorporate behaviour and preferences of stakeholders in the algorithmic decision-making process. The Software Engineering Research Group has a strong focus on empirical research, conducted in close collaboration with software development practice. SERG is known for its research on software testing, program analysis, mining software repositories, and empirical software engineering with close collaboration with industry partners.
As a candidate, you will have the opportunity to collaborate with experienced researchers, attend conferences, and experience an international research environment. The position will focus on research, with optional contributions to education and supervision of students.

JOB REQUIREMENTS

• You have a master’s degree in computer science, artificial intelligence, or mathematics.
• You can demonstrate an interest in computer security.
• In addition to a curious mind, you possess excellent programming skills.
• You have excellent communication skills in English, both written and oral.

Please refer the Job description for details