Sonic GAN
With Isart Digital Paris: adapted TOAD‑GAN to generate 100+ 2D Sonic‑style levels and integrated them into Unity.
Third-year Master’s student in Artificial Intelligence and Data Science at aivancity Paris-Cachan. Eager to discuss and participate in new projects. Let’s connect and build something amazing together!
Developed a zero‑shot ID document classification pipeline (90% accuracy on ~600 samples), benchmarked models and thresholds, and built a Streamlit interface that reduced review time by 40%. Wrote clear documentation and collaborated in Agile sprints.
With Isart Digital Paris: adapted TOAD‑GAN to generate 100+ 2D Sonic‑style levels and integrated them into Unity.
With X&Y Corp: Semantic grant matching system for associations using local LLMs