We are seeking to fill one position to undertake directed research activity as part of an EU project called FRAME (A Model-Based Foundry for Engineering, Adapting, and Assuring the Quality of AI Agents) under the direction of Prof Hamid Bouchachia. The project will deliver the engineering abstractions, methods and tools needed to support reliable, scalable end-to-end development, deployment and evolution of AI agents. By establishing foundations for trustworthy agent engineering, FRAME will enable productivity, adaptability, self-improvement, reliability and foster large-scale adoption. FRAME’s will be validated on real-world use cases (robotics, healthcare and software development frameworks).
The research associate will be affiliated with the Data Science & Intelligent Systems based in the Computing & Engineering Department. The group is very dynamic, ambitious, well networked and delivers state-of-the-art research in a range of machine learning and data science topics, publishing research results in prestigious venues.
We are looking for an individual with a talented and enthusiastic post-doctoral researcher who will join our team working on next-generation agentic systems capable of adapting to dynamic environments. The research work focuses on understanding and addressing the challenge of evolving real-world data over time and open-end learning in the context of self-adaptive and self-improving AI agents contributing to the development of methods for monitoring model behaviour, continuous evaluation, and improvement of agents across different training stages (e.g., pre-training and fine-tuning).
Depending on background, the applicant will pick one of the following topics:
- Developing novel indicators to capture distribution shifts, forgetting, transfer loss, etc. to monitor agent’s model performance and designing methods to recommend self-improvement strategies based on such system performance indicators.
- Evaluation of the effectiveness of different adaptation strategies under varying conditions, investigating and leveraging meta-learning and knowledge-driven approaches to guide dynamic selection of self-improvement techniques in response to evolving system requirements.
- Development and evaluation of continual learning techniques (e.g., continual pre-training, domain-adaptive pre-training, fine-tuning) and exploration of adaptive mechanisms for rapid adjustment, strategic retraining, and modular updates without full system redesign
- This is an excellent opportunity to gain hands-on experience in, among other topics, evolving agentic systems, continual learning, model robustness, and adaptive AI systems. The applicant should be holder of a PhD in artificial intelligence/machine learning and possess excellent background in agentic systems, large language models and/or continual learning besides analytical, programming, communication and scientific writing skills contributing effectively and competently to the delivery of
FRAME by designing and conducting the proposed research and producing published outputs.
The successful candidate will have access to funding for international travel, e.g., for attending conferences, attending the consortium meetings, and research dissemination, while working in a supportive, collaborative, inclusive and non-discriminating working environment. We welcome candidates from all backgrounds to apply
Please append to your CV (including publication list) and your formal application form, a max 2-page motivation letter to explain how your profile fits the open position.
For further information or discussion, please contact Hamid Bouchachia - [email protected]