Guillaume
Gillet

Full-featured web marketplace enabling buy/sell across categories. User account management, image uploads, search, and buyer–seller messaging.

Multi-part project building and benchmarking a high-performance server through iterative optimizations targeting maximum throughput and minimal latency.

Complete compiler frontend: lexer, parser, and semantic analyzer. Built a functional interpreter for a custom language with variables, procedures, and control flow.

Digital Captain Sonar in the Oz language. Multiple AI strategies, turn-based and simultaneous modes, sonar detection, mine laying, and missile mechanics.

Binary exploitation and malware analysis at the systems level. Covered multiple vulnerability classes, exploitation techniques, and static detection — hands-on offensive and defensive security.

Frequent itemset mining and sequential pattern mining. Implemented and benchmarked multiple algorithms across datasets for execution time, memory, and accuracy.

Cloud-native web app deployed to Microsoft Azure. Microservices architecture, containerisation, and distributed data processing showcasing modern cloud engineering.

10-year heart failure risk prediction using regression techniques and feature engineering on 800 medical records with 20 features, blending clinical data analysis with rigorous ML methodology.

Network security lab simulating a corporate topology in Mininet. Demonstrates multiple attack vectors and corresponding defensive countermeasures through live implementation.

Music-driven circle-based rhythm game built in Python. Sound effects, background music, and graphical interactions create an immersive audio-visual experience.

Full portfolio website built for ESILE, a Belgian singer-songwriter. Bio, music, photo gallery, live dates, media coverage, and booking form — pure HTML, CSS, and JavaScript, no framework.

Automatic medical entity recognition and SNOMED CT code linking from clinical notes. Four methods evaluated — dictionary matching, SapBERT bi-encoder, LLM NER+RAG, and a hybrid approach — on English (MIMIC-IV) and French clinical datasets.