Postdoc position on Federated/Continual Learning for Time-Series IoT Data (TRUMAN Project)

In this role, you will address the intricate challenge of enabling AI to learn continuously and collaboratively from wearable or mobile sensor data without compromising user privacy. Your efforts and collaborations with other European Union partners will contribute to the advancement of privacy-preserving artificial intelligence for the benefit of humanity.

What You Will Do:

• Research (Federated Continual Learning): You will develop novel and privacy-preserving algorithms that allow distributed devices (smartphones, wearables) to learn from new data streams over time (Continual Learning) while collaborating globally (Federated Learning).
• Analyze Mobile & Wearable Data: You will work with noisy, high-frequency time-series sensor data (e.g., accelerometers, gyroscopes, biosignals) from mobile IoT devices. You will address specific challenges such as user heterogeneity, concept drift in human behavior, and the need for personalization.
• Resource-Efficient AI: You will ensure your models are lightweight enough to run on edge devices, balancing accuracy with energy consumption and computational constraints.
• Project Coordination: You will assist the PI in managing work package tasks. This includes tracking technical progress, coordinating with partners on use-case requirements, and contributing to project deliverables.

About the TRUMAN Project:
The goal of TRUMAN is to design and develop generic technologies and methodologies for improving AI systems’ resilience against security, privacy, and fairness attacks, as well as to increase the trust that their users have in these systems, while accounting for different phases of the AI life cycle, starting from data collection through training and deployment. Your specific focus (Work Package 3) is on Human-In-The-Loop AI and Continual Learning. We aim to build systems that excel in performance but prioritize robustness, fairness, and privacy, critical requirements when dealing with time-series data. More info about the TRUMAN project can be found here.

We are looking for a researcher who sits at the intersection of Pervasive/Mobile Computing and Machine Learning.

• Background: You hold a PhD in Computer Science, Machine Learning, Electrical Engineering, Embedded Systems or related fields.
• Core Expertise:
• Strong expertise in Federated Learning and/or Continual Learning, with a solid understanding of the challenges involved in combining the two paradigms (FCL).
• Demonstrated ability to translate theoretical methods into working prototypes or deployable systems.
• Data and System
• Proven track record analyzing time-series IoT sensor data from wearables or mobile devices (e.g., human activity recognition, health and wellness monitoring, behavioral sensing).
• Comfortable working with imperfect, noisy, heterogeneous sensor data, and applying robust preprocessing, feature extraction, and validation workflows.
• Solid understanding of the practical constraints of mobile and embedded hardware (battery limitations, intermittent connectivity, limited compute/storage).
• Experience implementing or optimizing models for on-device or edge execution is a strong plus.
• Project Organizational Skills: You are well-organized and capable of coordinating a research work package, including planning, documentation, reporting, and partner communication within a collaborative project.
• Communication:
• Excellent communication skills in English.
• Comfortable working in interdisciplinary and international teams, engaging with both academic and industrial partners.

You will join the Pervasive Systems research group within the Department of Computer Science. The Pervasive Systems research group envisions a sustainable world achieved through the seamless integration of computing and the physical environment. We design collaborative, embedded sensing systems that interact unobtrusively with their surroundings. Our research focuses on developing scalable, adaptable, and trustworthy systems that ensure reliable, fault-tolerant performance despite inherent resource constraints. You will work in a social and multicultural research group where independence and personal growth are valued, but where you will never be alone.

Are you interested in this position? Please send your application via the 'Apply now' button below before February 16, 2026, and include:

• CV (including publication list)
• A cover letter highlighting your experience with time-series sensor data and distributed and or continual learning paradigms
• Research Statement.
• Two reference letters (one from your PhD supervisor).

Please contact Dr. Özlem Durmaz Incel if you have any additional questions via the following email adress: ozlem.durmaz@utwente.nl.