Portfolio

Ilias Laoukili — AI & Cybersecurity Engineer Student

Engineering intelligence and security through data, algorithms, and systems. Currently seeking 2025 research internships focused on trustworthy AI and security, open to relocation across the EU and US.

Profile photo of Ilias Laoukili

About

I'm an engineering student specializing in Artificial Intelligence and Cybersecurity. I like building fast, accessible interfaces and applying machine learning for practical problems such as anomaly detection and graph learning.

AI Focus

  • Graph Neural Networks
  • Model Optimization
  • Computer Vision

Security Interests

  • Threat Intelligence
  • Secure Systems
  • Ethical Hacking

Current Project — Tremplin Recherche (GNN Sparsification)

Research on Graph Neural Networks sparsification to improve inference latency and memory while preserving accuracy. Exploring structured pruning, topology-aware sparsity, and reproducible benchmarks with PyTorch Geometric/DGL.

  • Frameworks: PyTorch Geometric, DGL
  • Focus: latency · memory · accuracy
  • Deliverables: survey, experiments, demo

Featured Projects

Selected work showcasing AI and security applications

Stylized graph neural network with message passing flow highlighted.

Tremplin Recherche — GNN Sparsification

Structured sparsity experiments to accelerate GNN inference without sacrificing accuracy.

PyTorch Geometric DGL Model Compression
Drone vision interface highlighting fire and smoke detections with bounding boxes.

Drone Fire + Smoke Detection

Fine-tuned YOLO pipeline for onboard detection, cutting false alarms by 30% on field data.

Python YOLOv8 Edge AI
Echoes of Innocence Java game screenshot showing 2D gameplay mockup.

Echoes of Innocence — 2D Story Game

Narrative-driven Java adventure following Mathis in a conflict zone; built on clean OOP architecture.

Java OOP Game Design
Read project brief

Echoes of Innocence follows Mathis, a 10-year-old navigating a war-torn city to find his sister Lina. The game blends exploration, moral dilemmas, stealth, and inventory mechanics across locations such as a hidden rebel camp, a ruined cathedral, and the kidnapper’s lair—each revealing fragments of family memories.

Co-created with David J. Barnes, Michael Kolling, and D. Bureau, the project expanded the BlueJ world by refactoring navigation to HashMap exits, adding item collections with weight limits, and introducing branching endings driven by player choices.

Highlights

  1. 2025 — Present

    Volunteer, French Red Cross

    Coordinating logistics for outreach operations supporting 200+ families while streamlining supply workflows.

  2. 2021 — Present

    Academic Tutor & Instructor

    Guided 45+ students in math, physics, and CS with customized curricula, raising performance metrics by 15% on average.

Skills

Python
PyTorch
TensorFlow
Scikit-learn
Data Analysis
Deep Learning
Cybersecurity
Linux
Git

Get In Touch

Open to collaborations and interesting research/engineering challenges. Currently available for:

Research Projects Security Consulting AI Development

CV and project PDFs available above.