A PhD position is available through the InnoGuard MSCA Doctoral Network Project. Project 5 (Uncertainty quantification and handling of large language models for ACPS quality engineering tasks) and more information at innoguard.eu. Deadline to apply is November 1: application form.
The Innoguard project and Simula Research Laboratory are seeking a new PhD candidate to investigate how uncertainty in AI models can be quantified and used when AI supports quality engineering for autonomous cyber-physical systems (ACPSs). It covers AI models, including large models, that generate or assess artifacts such as test cases. The central goal is to determine when these outputs can be trusted and how uncertainty can inform risk-based engineering decisions. The doctoral candidate will develop metrics and methods, implement research prototypes, and evaluate them using open-source ACPSs, industrial case studies, and the InnoGuard industrial test and demonstration platform. The candidate will work with researchers across the network, contribute to joint training and dissemination activities, and complete the planned secondment at Mondragon Unibertsitatea.
The project Objectives are:
1) To develop new metrics for uncertainty quantification.
2) To develop methods that use quantified uncertainty to assess the risks and trustworthiness of AI models in ACPS quality engineering.
3) To study how uncertainty relates to correctness and non-functional characteristics, such as performance, of AI-generated quality engineering artifacts including test cases.
4) Investigate the applicability of the developed uncertainty metrics to advanced and future AI models, including quantum machine learning models (with the final phase of the PhD supported by an adjacent project).
Elegibility Criteria
Candidates of any nationality may apply. To be eligible for recruitment, an applicant must meet all of the following criteria on the recruitment date.
1. Doctoral candidate status The applicant must not hold a doctoral degree. An applicant who has successfully defended a doctoral thesis but has not yet formally received the degree is not eligible.
2. MSCA mobility rule The applicant must not have resided or carried out their main activity, such as work or study, in Norway for more than 12 months during the 36 months immediately before recruitment. Compulsory national service, short stays such as holidays, and time spent in a procedure for obtaining refugee status under the Geneva Convention are not counted.
3. Academic qualification By the recruitment date, the applicant must hold a master's degree or an equivalent qualification that gives access to doctoral studies in computer science, software engineering, artificial intelligence, machine learning, data science, robotics, or a closely related field.
4. English proficiency The applicant must have sufficient written and spoken English for doctoral research, publication, collaboration, and training.
Preferred profile
• Background in software engineering, artificial intelligence, machine learning, software testing, cyber-physical systems, or uncertainty quantification.
• Experience with Python or comparable programming languages and with empirical or experimental research.
• Interest in trustworthy AI and the evaluation of large AI models in safety-relevant systems.
• Ability to work independently and to collaborate in an international and interdisciplinary research network.
Application documents
Every application must include all three documents listed below.
1. Curriculum vitae Include education, relevant research or professional experience, publications if applicable, projects, and technical skills.
2. Academic transcripts Provide transcripts for completed and ongoing university degrees and include the grading scale if it is not shown on the transcript.
3. Motivation letter Explain your interest in the position, your fit with the research topic, and the experience and skills you would bring to the project.
Applications that omit any of these three documents may be treated as incomplete.
