The conference “Mathematics for Machine Learning in Burgundy” is an international meeting point for applied mathematicians, machine learning specialists, and AI researchers from a broad range of domains. Its goal is to strengthen the dialogue between rigorous mathematical foundations and the algorithmic, empirical, and application-driven perspectives that shape modern AI. The program features research and interdisciplinary talks, as well as dedicated networking moments (including a poster session) designed to encourage exchanges between communities. Particular emphasis is placed on connecting core mathematical tools, such as statistics, optimization, optimal control, and PDE-inspired modeling, with current challenges in machine learning, such as robustness, generalization, interpretability, etc. By bringing together academia and industry, the conference aims to spark new collaborations and accelerate the transfer of ideas from theory to impactful AI applications.