Nour L.

• Developed AI and statistical methods for continuous mood disorder monitoring using multimodal real-world signals. • Built end-to-end multimodal pipelines combining facial signals, GPS mobility, smartphone usage, and voice data for clinical outcome modeling. • Achieved correlations up to 0.67 between digital biomarkers and clinical scores. • Led applied R&D work with clinical impact, contributing to a patent-pending innovation on facial emotion indicators for mood disorder diagnostics. • Contributed to a publication validating the multimodal framework and its clinical relevance. • Developed models for irregular time series, including GRU-D, Random Forest, XGBoost, and GP-VAE for representation learning and imputation.

• Developed AI methods for continuous monitoring of mood disorders using multimodal real-world signals. • Built multimodal pipelines integrating facial, mobility, smartphone usage, and voice data for clinical outcome prediction. • Identified clinically relevant digital biomarkers and evaluated their association with validated clinical assessment scores. • Contributed to R&D work on multimodal digital biomarkers and clinical AI applications. • Modeled irregular and partially observed time series using GRU-D and GP-VAE-based approaches for imputation and prediction.

• Build anomaly detection models to identify behavioral disruptions in mood tracking data.

Built anomaly detection models for behavioral disruption in mood tracking data.

Developed EUR/ USD Forex predictive model.

Develop a predictive model for the EUR/USD foreign exchange rate.