PhD Candidate in Trustworthy Agentic Artificial Intelligence

This is NTNU NTNU is a broad-based university with a technical-scientific profile and a focus in professional education. The university is located in three cities with headquarters in Trondheim. At NTNU, 9,000 employees and 43,000 students work to create knowledge for a better world. You will find more information about working at NTNU and the application process here.     Video: https://www.youtube.com/watch?v=_KHQjc4ndas&t=41s About the position Would you like to work at the intersection of artificial intelligence, autonomy and maritime technology, contributing to AI systems that are safe, transparent and accountable? The Department of ICT and Natural Sciences at NTNU in Ålesund has a vacancy for a PhD Candidate in trustworthy agentic artificial intelligence. The PhD research project is part of the new Norwegian Maritime AI Center , led by NTNU. Agentic AI systems are moving rapidly from research prototypes into real-world applications. In safety-critical domains such as maritime operations, their adoption depends on whether their behaviour and decisions can be understood, verified, constrained, and audited. This is particularly challenging when using closed, proprietary models, where model weights, training data, and internal representations are inaccessible. The PhD project will therefore investigate how trustworthy agentic AI systems can be developed using open-weight models. Access to model weights enables the candidate to inspect and adapt the models, investigate their internal representations, apply interpretability and verification methods, fine-tune and align them for specific operational domains, and deploy them locally under full control of sensitive operational data. A central maritime use case will be condition monitoring and predictive maintenance of ships and maritime systems. Modern vessels generate large amounts of heterogeneous operational data from sensors, machinery, control systems, maintenance records, and other sources. The candidate will investigate how agentic AI can autonomously interact with such data and relevant engineering tools to identify abnormal behaviour, reason about possible faults, assess equipment condition, and support maintenance decisions. For example, an AI agent could combine sensor measurements and historical maintenance information, invoke diagnostic or predictive models, identify evidence of component degradation, and recommend appropriate maintenance actions. The position offers the opportunity to conduct cutting-edge research in trustworthy and agentic AI within a growing AI research environment, with access to relevant maritime expertise, data, and use cases and strong links to the Norwegian maritime industry. Your immediate leader will be the Head of Department.   About the project The center shall position the maritime sector for accelerated and successful use of AI. The demand for safe and sustainable maritime activities, as well as the shortage of crew and engineers, are key drivers for change in the maritime sector. At the same time, there is a pull for growth, increased efficiency and profitability, while managing volatility in the markets. Moreover, the sector is critically important for global supply chains and security in a tense geopolitical situation. To demonstrate and prototype maritime AI solutions, the center will address a diverse and impactful set of use cases: Logistics, shipping and business transactions; Ship design and construction; Voyage planning and optimization; Condition monitoring and maintenance; Crew training; Maritime traffic and ship surveillance; Decarbonization and energy management; Autonomous ships and remote control. The research partners are the Norwegian University of Science and Technology (NTNU), University of Oslo, Oslo School of Architecture and Design, Royal Norwegian Naval Academy, Norwegian Mapping Authority, MET and SINTEF Nordvest. There are about 30 user partners from the maritime sector, in addition to hundreds of members of the Maritime Forum, Blue Maritime, Ocean Autonomy, and Maritime CleanTech clusters. The center will be a hub, orchestrating the interfaces with research partners and user partners to facilitate collaboration, knowledge creation, dissemination, innovation, and sharing of knowledge. The Norwegian Maritime AI Center is funded by the Research Council of Norway and the partners for the period 2026-2030. There will be approximately 25 doctoral scholarships and some postdoc positions. More information can be found here: Norwegian Maritime AI Center.   Duties of the position • Conduct research on trustworthy agentic AI built on open-weight foundation models, including agent architectures, reasoning, planning and tool use, and the efficient adaptation of open models for locally deployed, domain-specific use. • Develop and evaluate methods that make agent behaviour reliable and accountable, such as uncertainty quantification, interpretability, alignment, monitoring and human oversight, taking into account relevant standards and regulation, including the EU AI Act. • Validate the developed methods on realistic case studies in maritime and other safety-relevant applications, in collaboration with the Maritime AI Center and industry partners. • Carry out research of good quality within the framework described above, including academic publications and popular science dissemination. • Participate in the research group Cyber-Physical Systems Laboratory (CPS Lab) and Norwegian Maritime AI Centre activities. • Take courses in relevant technical topics, comprising a minimum of 30 ECTS.   Required selection criteria • You must have a relevant Master's degree in computer science, artificial intelligence, data science, engineering cybernetics, applied mathematics, electrical engineering, or equivalent. Your course of study must correspond to a five-year Norwegian course, where 120 credits have been obtained at master's level. Master students can apply, but the master's degree must be obtained and documented before starting the position. • You must have a strong academic background from your previous studies and have an average grade from your Master's degree study, or equivalent education, which is equal to B or better compared to NTNU's grading scale . If you do not have letter grades from previous studies, you must have an equally good academic foundation. If you have a weaker grade background, you may be considered if you can document that you are particularly suitable for a PhD education. • You must meet the requirements for admission to the faculty's doctoral programme in Engineering Cybernetics . • Strong programming skills, in particular Python, and practical experience with modern machine learning frameworks. • Applicants from non-English speaking countries outside Europe must present an official language test report. The acceptable tests are TOEFL, IELTS, and Cambridge Certificate in Advanced English (CAE) or Cambridge Certificate of Proficiency in English (CPE). Minimum scores are: • TOEFL: 600 / writing 4.5 (paper-based) or 92 / writing 22 (internet-based) • IELTS: 6.5, with no section lower than 5.5 (only Academic IELTS test accepted) • CAE/CPE: grade B or A.   The appointment is to be made in accordance with  NTNUs guidelines for recruitment positions for general criteria for the position. Preferred selection criteria • Experience with agentic AI systems, LLM-based agents, tool use, planning, or multi-agent systems. • Knowledge of trustworthy AI topics such as interpretability, explainability, uncertainty quantification, robustness, formal verification, safety, privacy or AI governance and regulation. • Familiarity with maritime, autonomous systems, or other safety-critical application domains. • Demonstrated research ability, e.g. through publications, thesis work, technical reports or open-source contributions. • Relevant industry or applied research experience.   Personal characteristics To complete a doctoral de