Mahmoud
Hassan Ibrahim

Engineered a production-grade real-time streaming pipeline ingesting cryptocurrency market data from public APIs. The architecture features multi-stage processing with Kafka topics, Spark Structured Streaming transformations, and Airflow orchestration — fully containerized with Docker.

Designed and trained a Long Short-Term Memory neural network for time-series forecasting of stock prices. Applied advanced feature engineering — technical indicators, rolling statistics, momentum signals — and hyperparameter optimization to achieve robust predictive performance on historical market data.

Built an AI-powered motion capture pipeline using OpenFace and OpenPose to extract facial expression data and full-body skeletal keypoints from video input. The extracted motion vectors were used downstream for AI animation synthesis and motion transfer between subjects.

Built a production-grade Airflow DAG designed to manage heavy concurrent workloads without infrastructure overload. Implemented Airflow Pools to cap parallel task slots, Priority Weights to guarantee high-priority jobs execute first, SQL operators, and Python callables — all wired into a fault-tolerant orchestration workflow.

End-to-end ETL pipeline extracting live cryptocurrency data from the CoinGecko API, applying Python transformations, and loading clean records into PostgreSQL. Built with production-style structure: environment variable security, structured logging, error handling, and modular code architecture.

Engineered an Airflow monitoring DAG that automatically detects task failures and SLA violations, then dispatches email alerts before deadlines are breached. Enables proactive incident response — shifting teams from reactive firefighting to proactive pipeline observability.

Designed a no-code workflow automation system using n8n for pharmacy inventory management. Automated stock tracking, low-inventory alerts, and inventory update workflows — eliminating manual data entry and reducing human error in pharmaceutical stock operations.

Built an AI-powered customer service automation workflow using n8n to handle common support queries, route escalations, and log interactions automatically. Reduced response times and freed human agents for complex cases by automating first-line customer interaction handling.

Developed a workflow automation system for library operations using n8n — automating book check-in/check-out tracking, overdue notifications, and catalog updates. Demonstrates ability to map complex operational logic into automated, reliable workflows without traditional coding overhead.