PhD Researcher in Trustworthy and Human-Centred AI Development

Descrizione

Testo originale dell'annuncio (in inglese)

PhD Researcher in Trustworthy and Human-Centred AI Development The KOF Swiss Economic Institute at the ETH Zurich is the leading institute for applied research in economics in Switzerland. We are a research group within the KOF Swiss Economic Institute at ETH Zürich. Our work sits at the intersection of economic research and AI: we study money systems and democracies, and we build AI systems that engage with these domains — for research, for policy, and the public. We are now looking for a PhD Researcher who combines strong engineering fundamentals with serious attention to how these systems behave in the world and who they serve. Job description As a PhD Researcher, you will pursue independent research on trustworthy and human-centred AI while contributing to the systems we are building in this space. The work spans methodological research and concrete implementation, and treats questions of reliability, accountability, and human-AI interaction as research questions in their own right and not as deployment afterthoughts. You will work alongside researchers and engineers in an international team and will have the room to shape the direction of your own thesis. An example of a system to be developed are trustworthy AI voting assistants in the context of Supported Democracy . Your responsibilities will include: Conducting research on AI reliability, security, trustworthiness, and human-AI interaction, with a focus on high-stakes settings. Publishing at top venues in AI and responsible AI research (NeurIPS, ICML, ICLR, ACL, FAccT, EAAMO, AIES, CHI, etc.). Designing, building, and iterating prototypes that take their real-world context seriously, including who uses them, under what conditions, and with what consequences. Working hands-on with large language models: evaluating their behavior, improving their reliability for concrete use cases, and integrating them into systems with particular attention to security, explainability, and the interaction between humans and AI. Contributing to the development of LLM-based systems, including prompting strategies, tool use, evaluation, retrieval-augmented workflows, and adaptation. Designing and running user studies to evaluate how people interact with the systems we build, in collaboration with other members of the team. Profile We are looking for someone who is technically strong and who is comfortable moving between research, engineering, and the broader questions of how AI systems affect the people and institutions that interact with them. You should bring: A master’s degree in computer science, machine learning, HCI, or a related field. Strong foundations in machine learning and familiarity with current AI tools and practices. A solid understanding of large language models, in particular their reliability, security, and explainability characteristics. Familiarity with the technical and human-factors sides of human–computer interaction. The ability to work independently, communicate clearly, and navigate open-ended research question The following would be a plus: Practical experience building with LLM APIs, open models, evaluation pipelines, orchestration frameworks, and related tooling. Familiarity with model adaptation, fine-tuning, or related optimization methods. Experience with cloud infrastructure, MLOps or LLMOps, containerization, or production deployment. Frontend and/or backend development skills. We offer ETH Zurich offers internationally competitive working conditions. The starting date for this position is August or September 2026 and the duration is initially 18 months with the prospect of a further extension up to a total of six years, depending on age and qualification. We offer the opportunity to pursue a high-quality PhD at renowned institutions in a dynamic and collaborative research environment. ETH Zurich has internationally competitive working conditions. The gross salary will be the standard salary rate for PhD students at ETH Zurich. KOF PhD students are part o