Summary
John Gkountouras is a PhD candidate in Artificial Intelligence at the University of Amsterdam with nine years of industry and research experience bridging deep learning, causality, and LLM grounding. A mathematician by training (BSc in pure math, dual minors in CS and computational math) and MSc (cum laude) in AI, he combines rigorous theory with hands-on engineering—proficient in Python/PyTorch, efficient LLM fine-tuning, quantization, and inference optimization. He has led data teams and product delivery, built end-to-end ML systems for real estate valuation, and interned on LLM compression research at Stochastic. Currently his research focuses on L(V)LMs as agents and causality-inspired world models that ground language in vision, reflecting a rare mix of formal logic, optimization obsession, and production-minded implementation.
9 years of coding experience
2 years of employment as a software developer
Bachelor of Science, Pure Mathematics, Computer Science, Computational Mathematics, Bachelor of Science, Pure Mathematics, Computer Science, Computational Mathematics at University of Ioannina
Doctor of Philosophy - PhD, Artificial Intelligence, Doctor of Philosophy - PhD, Artificial Intelligence at University of Amsterdam
English, French, Greek