title: “Projects” permalink: /projects/ author_profile: true —
Selected Projects
SAFE-ICU Data Resource
SAFE-ICU is one of the central outcomes of my PhD work and one of the most important milestones in my research profile. It brought together clinical data collection, ICU data pipelines, predictive modeling, and deployable AI systems for critical care research and decision support.
The SAFE-ICU ecosystem supported work on early warning systems, multimodal monitoring, and practical research software for intensive care environments. It also helped build a strong foundation for reproducible and translational healthcare AI research.
- Project page: SAFE-ICU
- Google Scholar profile: Publications
- Previous lab page: TavLab profile
ShoQPred
ShoQPred is a deployable human-centric software suite for hemodynamic shock prediction using physiological vitals time-series. The system was designed to support early warning in ICU settings and to translate predictive models into practical clinical tools. [web:85]
- GitHub: SAFE-ICU/ShoQPred
Automated Shock Detection using Thermal Imaging
This work explored non-contact monitoring and early warning through thermal imaging-based analysis. It is part of the SAFE-ICU research line and reflects my interest in multimodal AI for real-world healthcare environments. [web:93]
FAIR Healthcare Data Harmonization
A major part of my applied AI work has focused on improving the usability of real-world healthcare datasets. This included harmonizing non-curated healthcare data, improving interoperability, and building data pipelines that support downstream analytics and machine learning. [file:35]
Foundation Models for Biology Data
My current work at INRIA focuses more strongly on computer science and machine learning methods for biology data, especially small foundation models, representation learning, and practical model design for scientific use cases.
Selected Publications and Research Directions
Below are representative themes from my publications and project work. For the full and most up-to-date list, please see my Google Scholar profile. [file:35]
- Early prediction of hypothermia in pediatric intensive care units using machine learning. [web:81]
- Hemodynamic shock prediction and deployable ICU early warning systems linked to SAFE-ICU resources. [web:85]
- Thermal imaging-based analysis for automated shock detection in ICU environments. [web:93]
- FAIR and interoperable healthcare data pipelines for AI-ready datasets. [file:35]