Engineering Student in Data Science, Big Data & Artificial Intelligence

github

medium

MANAL ES-SOBHY  • Peerlist

ClearPath AI is a smart-city platform designed to optimize emergency vehicle response times through AI-driven traffic prediction and dynamic route management. The system leverages real-time and historical traffic data to create priority corridors for ambulances, fire trucks, and police vehicles, reducing delays during critical situations. Key Features: • Traffic congestion prediction using AI and data analytics • Real-time route optimization for emergency vehicles • Dynamic traffic signal coordination and green corridor creation • Integration with existing smart traffic infrastructure • Centralized coordination of multiple emergency services Impact: Designed to improve emergency response efficiency, reduce travel times for critical services, enhance urban mobility, and support smarter, safer cities through intelligent traffic management.

ClearPath AI - RallyIA 2026

Developed an AI-powered insurance callbot designed to automate first-level customer support requests, including claim follow-ups, policy information, and FAQs. The solution combines Generative AI, Agentic AI, Retrieval-Augmented Generation (RAG), and NLP techniques to provide accurate, context-aware responses while reducing agent workload. Key Features: • Natural language understanding and intent detection • Entity extraction and contextual information retrieval • RAG pipeline with semantic search over a Chroma vector database • Integration with business data sources for customer-specific information • Multilingual conversational support with voice input/output capabilities Project developed during the GenAI Morocco Hackathon organized by Capgemini Morocco.

Julie - GenAI & AgenticAI National Hackathon 

Designed Saha City, a Monopoly-style educational board game aimed at raising awareness about sexual and reproductive health (SRH) in a fun and accessible way. The game transforms sensitive and complex topics into interactive gameplay, using simple mechanics to make learning more engaging, accessible, and intuitive. It encourages players to explore concepts through discussion, decision-making, and participation in a safe and open environment. In addition to the gameplay design, a modern and inclusive visual identity was developed to support clarity, accessibility, and promote open and approachable communication.

Saha City - Digi-Hackathon of Mobile Health 2

Collaborated with a multidisciplinary team to develop TabooTalk, a digital solution designed to promote awareness and education around sensitive health topics, especially among youth. Key Features: • Interactive Chatbot (TT): One-on-one Q&A-based chatbot for personalized learning and support. • Knowledge Quiz: Single-choice questions (QCU) to test and reinforce users' understanding of core concepts. • Interactive Forum: A safe space for users to express their thoughts and experiences in text format, promoting community dialogue and peer support. This project focused on using AI and digital technology to enhance health education and break taboos around sensitive subjects.

TabooTalk – Digi-Hackathon of Mobile Health 2

Developed an end-to-end MLOps and DataOps pipeline for predictive maintenance, designed to predict the Remaining Useful Life (RUL) from sensor data. The solution integrates real-time data streaming, automated data processing, feature management, machine learning lifecycle management, and monitoring components to simulate an industrial AI maintenance platform. Key Features: • Real-time engine telemetry ingestion using Apache Kafka for continuous data streaming • Data processing pipeline with Bronze/Silver/Gold Data Lake architecture • Automated data cleaning, transformation, and feature engineering workflows • Feature Store implementation using Feast for centralized offline/online feature management • MLOps pipeline with MLflow for model training, experiment tracking, evaluation, and versioning • Deep learning-based RUL prediction model for predictive maintenance applications • Real-time monitoring dashboard using Streamlit for sensor visualization and maintenance insights • Industrial supervision interface for machine fleet monitoring, alerts, and operational analysis • Containerized deployment environment using Docker for reproducible execution