PhD Researcher in Learning AI Agents
Explore human-AI interaction and control mechanisms for autonomous learning agents in industrial settings.
Best for:Best for computer science graduates with expertise in human-AI interaction, reinforcement learning, and behavioral drift detection, seeking a multidisciplinary PhD opportunity at the intersection of AI, human factors, and organizational control.
Overview
This PhD position at the University of Luxembourg focuses on studying how organizations maintain meaningful control over AI agents that continuously learn from user interactions. The research addresses two core challenges: determining when an agent should proactively act versus waiting for explicit instructions, and detecting behavioral drift in agents to ensure their behavior remains aligned with validated baselines. The work combines theoretical research, system development, and empirical evaluation with real users in operational environments. The successful candidate will collaborate with academic and industrial partners, particularly a major Luxembourgish energy utility, and contribute to open-source software tools. The position is based at the UBIX Research Group within the Interdisciplinary Centre for Security, Reliability and Trust (SnT).
What you'll do
- Conduct research on human-AI interaction, focusing on how agents calibrate initiative and present choices without overwhelming users.
- Develop methods for detecting behavioral drift in learning AI agents and designing mechanisms to restore trusted behavior.
- Design and implement proof-of-concept software tools for AI agents, with a focus on open-sourcing where applicable.
- Conduct empirical studies with real users in operational industrial settings to evaluate AI agent performance.
- Collaborate closely with academic and industrial partners to disseminate research findings through publications and project activities.
- Present research results to project teams, industrial partners, and academic communities.
- Publish findings in scientific journals and conferences, contributing to the broader AI research community.
Requirements
Degree required: Master of Science (MSc) degree or equivalent in Computer Science or a related engineering discipline
Degree offered: PhD
Language: Excellent command of English; B2-level proficiency in the language of the thesis required
Instruction: English
- Strong analytical and programming skills in Python and C/C++
- Experience with large language models (LLMs) and LLM-based agents
- Knowledge of human-AI interaction, human-computer interaction, and human factors
- Familiarity with reinforcement learning and learning from human feedback
- Experience with behavioral and distribution drift detection in AI systems
- Background in statistical anomaly detection or evaluation/benchmarking of AI systems
- Applicants must hold a relevant master’s degree or equivalent
- No specific nationality or residency restrictions mentioned
- No age limits specified
Funding & terms
How to apply
Deadline: 2026-10-20
Applicants must submit their CV, a cover letter explaining their motivation and alignment with the research focus, transcripts of all university-level courses, and a link to a GitHub repository (if available) through the university’s online HR system. Early applications are encouraged as they will be processed upon reception. Email applications will not be considered.
- CV including degree details, thesis summary, list of publications, and GitHub repository link (if applicable)
- Cover letter detailing motivation and alignment with the research topic
- Transcript of all university-level courses and results
Contact
HR Department· Recruitment Contact
Create a free profile to unlock your personal match score, the full requirements, funding details and how to apply — in English or Persian.