At Vrije Universiteit Brussel, I worked on a novel distributed stereo-image compression architecture. I researched relevant topics and implemented and optimized architectures, data pre-processing pipelines, and AI development tools.
At NightOwl Studios, I implement and evaluate LLM-based chatbot assistants using retrieval-augmented generation. I work with gte-series embedding models, PostgreSQL, vector databases, and the Claude API.
I also develop safe API endpoints in .NET 10.0. Earlier, at Eunoia, I optimized database queries, engineered interfaces in React, and developed server-side APIs in .NET.
My work and studies have focused on classical and learning-based image compression, object detection, and computer graphics. My thesis was on task-oriented learned stereo-image compression, and my final-year project on optimizing architectures for Alzheimer's disease classification received the Best Final-Year Project Award.
