Andoni Irazusta Garmendia

I’m Andoni Irazusta Garmendia, a researcher and engineer working on deep learning systems that learn from experience, adapt at inference time, and act intelligently in complex decision-making settings.

A large part of my work uses combinatorial optimization problems (COPs) as a challenging testbed for these ideas. In that context, I develop end-to-end approaches that combine representation learning, search and decision-making, and reinforcement learning, and I build the full pipeline from model architectures and training objectives to evaluation and large-scale experimentation.

If you’d like a quick overview, check out my Blog and Publications pages, or reach out via email.

Outside research, I like running and experimenting with performance training.