Nouha Karaouli
I'm PhD Student
About
I’m Nouha Karaouli, an Embedded Systems Engineer currently pursuing my PhD at the University of Rennes, working at the IRISA Laboratory within the MALT team.
My doctoral research focuses on Incremental Learning for Embedded Time Series Foundation Models.
I am particularly interested in foundation models, especially those based on Transformer architectures, for time series forecasting. My research focuses on incrementally adapting these powerful models while addressing the challenge of catastrophic forgetting. We then aim to further adapt these models for efficient implementation on embedded systems, which operate under stringent constraints in terms of energy consumption, computational resources, and memory footprint.
- Starting on: December 2024
- City: Rennes, France
- Degree: PhD
- Email: nouha.karaouli@irisa.fr
Research Interests
Focus areas and topics that drive my PhD research and passion
Time Series Analysis
- Univariate & Multivariate Time Series
- Time Series Foundation Models
Incremental Learning Methods
- Regularization-based Approaches
- Parameter-Efficient Techniques
- Replay Strategies
Efficient Architectures
- Mixture of Experts (MoE)
- Distilled Models
- Model Compression & Optimization
Resume
Summary
Nouha Karaouli
PhD Candidate at Rennes University, IRISA Lab (MALT team). My research focuses on incremental learning for embedded time series foundation models. I leverage transformer-based models to enable efficient forecasting under embedded systems constraints (energy, computation, memory).
Education
PhD in Embedded Systems & Machine Learning
Dec 2024 – Present
Rennes University, IRISA Lab – MALT Team
National Engineering Degree in Infotronics
2021 – 2024
National Engineering School of Carthage
Preparatory Program in Physics and Chemistry
2019 – 2021
Faculty of Sciences of Tunis
High School Diploma, Science
2018 – 2019
Gazela City, Wafa High School
Professional Experience
substitute teacher - Machine Learning
Jan – Mar 2025
ISTIC, Rennes
- Taught introduction to machine learning models (linear/logistic regression, decision trees, neural networks, etc.) using the R language and supervised the final project of this course.
Research Intern
Apr - Oct 2024
Centralesupélec-IETR Lab, Rennes
- Enhanced physical layer security by developing a Machine Learning-based Device Identification System.
- Implemented deep learning models to analyze RF fingerprinting techniques, ensuring precise identification of transmitting devices.
- Contributed to network security improvements against unauthorized access.
Engineering Intern
Jul – Aug 2023
Bee Coders, Tunis
- Building a data-driven competitive analysis platform using social media scraping APIs, enabling optimized decision-making for the marketing team.
Contact
Feel free to reach out for questions or collaborations.
Address
IRISA UMR 6074Centre de recherche Inria de l'Université de Rennes
263 Avenue du Général Leclerc
35000 Rennes, FRANCE
Phone
+33 6 37 96 26 51
nouha.karaouli@irisa.fr