Abdelrahman Ewida — ML & Optimization Research

About

I work at the intersection of applied machine learning and systems engineering: building models that run against real sensors, real constraints, and real latency budgets rather than static benchmarks. My thesis at the Human-Centered AI lab, Universität Augsburg (Prof. Elisabeth André), produced a real-time wearable stress-detection and intervention system, completed with grade A+.

I am a Category "A" scholar at the GUC and a national-level competitive swimmer, which is where I learned that consistent, measured training beats sporadic intensity — in the pool and in research.

Publications

EAI PervasiveHealth 2025 · Springer LNICST

Design of Wearable-Based Real-Time Stress Detection and Breathing Intervention for Workplace Well-Being

Abdelrahman Ewida, Thomas Kiderle, Elisabeth André, Bhargavi Mahesh

Lead author. End-to-end system that streams ECG from a Polar H10 chest sensor over BLE, extracts HRV features in sliding windows, classifies stress with a Random Forest, and delivers a context-aware guided breathing intervention that defers to the user (no interruption while typing; user confirms before it starts).

81.05% mean accuracy under leave-one-subject-out cross-validation on WESAD (15 subjects), deployed in a live React Native / Flask pipeline at ~130 Hz.

[PDF on request] doi.org/10.1007/978-3-032-27585-1_27

ICEENG-IPCS 2026 · IEEE

Optimization of Multi-AGV Scheduling for Airport Baggage Handling

Abdallah A. Hassan, Abdelrahman Ewida, Haidy E. Elkenawy, Abdulrahman B. Bahy, Karim M. Fathy, Dalia M. Mahfouz, Omar M. Shehata

Co-author. Multi-objective VRPTW formulation for airport baggage AGVs, benchmarking Simulated Annealing, Genetic Algorithm, Ant Colony Optimization, and Q-Learning against a proposed Hybrid Artificial Bee Colony algorithm combining discrete routing operators with continuous start-time perturbation.

Hybrid ABC improved operational efficiency by ~61% over the FCFS baseline while cutting fleet size from 11 to 3 AGVs on the standard instance; Q-Learning traded optimality for ~1 s execution, motivating a hierarchical plan-then-recover architecture.

[PDF on request] doi.org/10.1109/IPCS69631.2026.11604463

Selected Projects

ClusterChaos: Amazon Microservices Replica

Java · Spring Boot · Kubernetes · RabbitMQ

A 5-service e-commerce backend (team of 13) with per-service polyglot persistence across PostgreSQL, Cassandra, Neo4j, Elasticsearch, and Redis, a choreography saga over RabbitMQ with DLQs and idempotent consumers, a JWT-validating API gateway, and full Kubernetes deployment with Prometheus, Grafana, and Loki observability.

github.com/Tatos23/clusterchaos-amazon

SBOM Generator

Python · Docker · Click · CycloneDX / SPDX

A command-line tool (built in a team of 4) that scans Python and Java projects, resolves transitive dependency trees via the PyPI and Maven Central APIs, generates schema-validated CycloneDX 1.7 and SPDX 3.0 SBOMs, and produces OSV-based CVE reports sorted by severity.

github.com/Tatos23/sbom-generator

Intelligent Airline Assistant & Knowledge Graph

Python · Neo4j · Scikit-Learn

Passenger-satisfaction prediction pipeline (Random Forest) with SHAP and LIME explainability, paired with a Neo4j knowledge-graph RAG system that answers complex travel queries through vector embeddings and graph traversal.

github.com/Tatos23/Airline-Customer-Holiday-Booking

AI Game Agents & Logic Reasoning

Java · Prolog · Minimax · Situation Calculus

Two agents: an adversarial search agent that plays a turn-based battle game with Minimax and alpha-beta pruning, and a Prolog planning agent that reasons in the Situation Calculus, using iterative deepening to keep backward chaining complete.

github.com/Tatos23/ai-game-agents

Custom Database Management System

Java · B+ Trees

A database engine built from scratch: B+ tree indexing, page serialization, and binary search over pages to support efficient CRUD operations — no libraries doing the heavy lifting.

github.com/Tatos23/databaseEngine

Ajaza: Virtual Trip Planner

MERN · MongoDB · Express · React · Node.js

A full-stack web application for travel budgeting and itinerary planning, integrating third-party APIs for real-time booking data and secure user authentication.

github.com/Tatos23/Ajaza

Experience

Research Thesis Student

— Human-Centered AI, Universität Augsburg, Germany 2025 · semester exchange

Wearable sensor integration for breathing-based emotion regulation, supervised by Prof. Dr. Elisabeth André. Resulted in the PervasiveHealth 2025 publication. Grade: A+.

IT Networks & Applications Intern

— National Automotive Company (NATCO), Egypt Jun – Aug 2024

Built an employee records management system in C#/.NET with EF Core; designed normalized SQL schemas and optimized reporting with LINQ.

Curriculum Vitae

Download CV (PDF)