Mariam Mchirgui

Developed a multimodal AI system to detect depression from text (Twitter data), speech (clinical interview recordings), and EEG signals.

Designed late-fusion architectures combining deep models and evaluated performance on imbalanced clinical data.

Achieved high screening accuracy and reduced false negatives for depressive subjects.

Work accepted for presentation at IEEE AMCAI 2025.

Designed and developed a comprehensive Human Resources portal
using Power Apps, Power Automate and SharePoint Lists.
Created a Check-in/Checkout app to help employees track
attendance.
Developed a Kudos app to facilitate employee recognition.
Built a Leave Requests app to automate leave management and
approval processes.
Implemented a Document Workflow app to automate document
approvals and version control.

Implementation a management and monitoring of TT infrastructure
and installation of wireless technologies and fiber optic 4G/5G
networks.
Integration with OpenStreetMap for mapping and visualizing MSAN and
future GPON locations
Designed Interactive dashboard for zooming, panning, and analyzing
geographic locations of network nodes

Developed Python-based medical self-assessment platform for
symptom-based sickness prediction
Built machine learning models for classifying symptoms and
suggesting potential diagnoses
Performed data preprocessing and feature engineering on healthcare
datasets
Collaborated on platform design, user interface, and backend
integration