At TÜBİTAK, I designed and deployed computer vision, deep learning, and multimodal AI solutions for national public-sector projects across defense, broadcasting, and retail. For Market Fiyatı, I developed AI validation using Vision–Language Models and GPT-based LLMs, reducing manual review effort by over 50% through confidence-based automation and human-in-the-loop validation.
I also designed and operated a national deepfake detection competition and built a test dataset for a system that produces authenticity verdicts. In a personal project, I built and deployed an AI meeting platform that turns recordings and documents into searchable, source-cited knowledge.
