Loïc Guibert - PhD Student in Data Science

2025.08.06, ACM SIGENERGY Energy Informatics Review @ Cambridge, MA, USA.

Authors: Olivier Weppe, David Bekri, Thibaut Marty, Loïc Guibert, Louise Aubet, Jean-Christophe Prévotet, Maxime Pelcat, Sébastien Rumley

2025.07.06, CF '25 Companion: Proceedings of the 22nd ACM International Conference on Computing Frontiers @ Cagliari, Italy.

Authors: Loïc Guibert, Sébastien Reynaud, Olivier Weppe, Kristjon Ciko, Michael Welzl, Sébastien Rumley

2020.11.02, International Conferences on Internet of Things @ Rhodes, Greece.

Authors: Pascal Bruegger, Adriana Wilde, Loïc Guibert

2024.02.06, AI-Days @ Lausanne, Switzerland.

Authors: Loïc Guibert, Pascal Bruegger

Development of an Open-Source Kit to facilitate the publication of research software as open source at HES-SO, with a real-case implementation.

Duration of 3 months, finished the 2025.07.01.

Carried out @ School of Engineering and Architecture of Fribourg, managed by Sébastien Rumley. Funded by HES-SO.

This project aims to develop a REST service allowing to classify documents into categories and to extract their entities using large language models (LLMs).

Duration of 2 months, finished the 2025.06.30.

Carried out @ School of Engineering and Architecture of Fribourg, managed by Sébastien Rumley. Funded by Swiss AI Center.

The EIAEAE project aims to develop a method enabling companies to assess the harmful impacts caused by their use of artificial intelligence.

Duration of 10 months, finished the 2024.12.01.

Carried out @ School of Engineering and Architecture of Fribourg, managed by Sébastien Rumley. Funded by HES-SO.

A collaborative project between technical and management schools to support entrepreneurship.

Duration of 1 year, finished the 2024.02.01.

Carried out @ School of Engineering and Architecture of Fribourg, managed by Pascal Bruegger. Funded by HES-SO.

Hestia brings technological resources within nursing homes, personal homes or protected apartment in order to improve the life comfort of vulnerable people and the working conditions of their medical staff.

Duration of , finished the 2023.01.01.

Carried out @ School of Engineering and Architecture of Fribourg, managed by Pascal Bruegger. Funded by Nouvelle Politique Régionale (NPR), Fribourg Canton and HES-SO.

RMS brings technological resources within nursing homes, personal homes or protected apartment in order to improve the life comfort of vulnerable people and the working conditions of their medical staff.

Duration of 5 years, finished the 2022.12.31.

Carried out @ School of Engineering and Architecture of Fribourg, managed by Pascal Bruegger. Funded by HES-SO.

2023-2024 and 2021-2022. Managed by Pascal Bruegger, Sandy Ingam and Michael Mäder.

This lecture compots minimal theorical lectures, and focuses on practical work. Groups of students will learn how to develop web clients, mobile applications, and backends using modern frameworks and tools. Communication and organisation skills are also covered and evaluated alongside the technical skills.

2020-2021. Managed by Pascal Bruegger.

This lecture extend knowledge aquired on Android developemnt given a semester before. iOS developpment basics are covered, in order to then teach about cross-platform frameworks. A large space is given on practical works.

2023-2024. Managed by Pascal Bruegger and Aïcha Rizzotti.

This module allows students to develop their knowledge of multi-platform and cross-platform mobile application development (Android and iOS) in terms of both logic and graphical interfaces. Using platforms such as Flutter, Xamarin and Ionic, students become familiar with best practices in mobile development. As part of a mini-project, students work on both the logic of the app and the creation and design of user-friendly graphical interfaces, applying the concepts of user experience and testing learned in class.

2020-2021. Managed by Pascal Bruegger.

This lecture covers the fundamentals of mobile application development, focusing on the Android system. Students will learn about user interface design, data management, and integration with backend services. Practical sessions will involve building a complete mobile application from scratch, applying best practices in coding, testing, and deployment.

Duration of 8 weeks, finished the 2026.03.19.

This project transformed TRAFFIC from a bloated 83,000-line monolith into a sleek, plugin-based architecture that slashed the codebase by over 95% while automating complex deployments. Version 2.0 makes measuring data center energy consumption significantly more extensible and user-friendly without compromising performance or accuracy. It's essentially a high-performance "diet" for a tool that was previously weighing itself down with massive technical debt.

Duration of 1 semester, finished the 2026.03.04.

This thesis investigates the trade-off between accuracy and energy consumption in open-source Document AI solutions. Key findings reveal no direct correlation between energy usage and performance, challenging traditional scaling laws. The study identifies PaddleOCR VL expert models as optimal for complex documents and Vision Language Models as the most sustainable architecture. To assist users, the thesis concludes by proposing a decision tree for selecting the best Document AI solution based on specific requirements.

Duration of 1 semester, finished the 2026.02.03.

This study assesses integrating AI into Hestia, a home medical monitoring platform using physiological time series. Experiments revealed moderate performance and insufficient recall due to measurement limitations, making immediate clinical use infeasible. Nevertheless, the research concludes that AI integration remains plausible provided robust data and reliable clinical labels are available.

Duration of 8 weeks, finished the 2024.07.14.

This project focuses on integrating ECG (electrocardiogram) monitoring into the Hestia platform, a digital health ecosystem that streamlines medical data sharing between patients and doctors, to enhance the detection and management of cardiac conditions like arrhythmias and myocardial inflammation.

Duration of 1 semester, finished the 2024.05.16.

This student project focuses on developing a module for the Hestia platform to automate the direct transfer of physiological data from various smartwatches to doctors, providing a seamless monitoring solution that functions autonomously without depending on third-party manufacturer apps.

Duration of 8 weeks, finished the 2021.07.16.

This project addresses the challenges of elderly home care and overwhelmed medical facilities by leveraging technologies to improve medical response times and provides a monitoring platform suitable for healthcare professionals, families, and athletes.

Duration of 8 weeks, finished the 2021.07.16.

Dutine is a cross-platform mobile application designed to address the challenges of tracking repetitive tasks and maintaining long-term user motivation. By implementing a gamified system where 'givers' assign tasks and 'receivers' earn real-life rewards or face punishments, the project moves beyond traditional virtual badges to impact the user's daily life. Developed using Flutter and Firebase, the thesis covers the entire software development lifecycle, including a comprehensive analysis of user needs, competitor products, and gamification principles, culminating in a functional app successfully deployed on both the App Store and Play Store.

Duration of 8 weeks, finished the 2021.07.16.

This project involves the improvement and securisation of a web-based appointment booking system integrated with existing management software, on the basis of a proof of concept developed during a previous semester project.

Duration of 1 semester, finished the 2021.05.26.

This project aims to develop a secure, web-based calendar allowing physiotherapy patients to book and manage appointments online using a unique identification number.

Duration of 1 semester, finished the 2021.02.03.

The in-house RMMA application monitors patient vitals via a chest strap. To improve patient comfort, this project replaces chest straps with smartwatches by updating the RMMA application to collect and display heart rate, blood pressure, and accelerometer data via Bluetooth.