ScaDS.AI Dresden/Leipzig announces and welcomes you to join its public colloquium session on Wednesday, December 17, 2025 at 11:00 am CET. The colloquium takes place at the Felix Klein lecture hall (P 501) in Leipzig University’s Paulinum building (details below) and parallel online (link to Zoom session). Also note his later talk in presence at 4:00 pm in the Felix Klein Colloquium of the Institute of Mathematics, titled “Discovery of unstable singularities”.

AlphaEvolve: Exploration and discovery at scale

AlphaEvolve is a generic evolutionary coding agent that combines the generative capabilities of LLMs with automated evaluation in an iterative evolutionary framework that proposes, tests, and refines algorithmic solutions to challenging scientific and practical problems.

In this talk we will present AlphaEvolve, discussing its implications as a tool for autonomously discovering novel mathematical constructions and algorithms, advancing our understanding of long-standing open problems. Moreover, AlphaEvolve is also able to generalize results for a finite number of input values into a formula valid for all input values. And in some instances we were able to combine this methodology with Deep Think and AlphaProof in a broader framework where the additional proof-assistants and reasoning systems provide automated proof generation. These results demonstrate that large language model-guided evolutionary search can autonomously discover constructions that complement human intuition, at times matching or even improving the best known results, highlighting the potential for significant new ways of interaction between scientists and AI systems.

We present AlphaEvolve as a powerful new tool for discovery, capable of exploring vast search spaces of algorithms to solve complex optimization problems at scale, often with significantly reduced requirements on preparation and computation time.

About Javier Gómez Serrano

His research interests lie at the boundary between AI and mathematics. His work has been covered by scientific and general media such as Quanta Magazine, Quanta Podcast, Communications of the ACM, Spektrum, or El País, highlighting his pioneering use of artificial intelligence in solving fundamental mathematical problems.

He has given over 150 invited talks at conferences and seminars at venues such as Princeton, MIT, Harvard, Yale, NYU Courant, UCLA, Columbia, IAS, Simons Foundation, National Academy of Sciences, Cambridge, Oxford, ENS… or distinguished lectures such as the Riviere-Fabes Symposium (2025) or the Harvard Millennium Prize Lectures (2026).

Javier has been awarded many distinctions, highlighting: the Ramon E. Moore Prize (2025), the MCA Prize (2025), the Antonio Ambrosetti Medal (2023), the Antonio Valle Award (2018), the Vicent Caselles Award (2017), and Princeton’s Junior Faculty Teaching Award (2015). He is a Simons Fellow (2024) and an IAS Member (2025), and has held an Aisenstadt Chair (2024).

Born in Madrid in 1985, he received a PhD from Universidad Autónoma de Madrid in 2013 (best thesis of the year). He then moved to Princeton University as an Instructor (2013-2016) and later as Assistant Professor and Director of Graduate Studies (2016-2020). Between 2020 and 2022, Javier held a shared position between the University of Barcelona (as Distinguished Researcher / Full Professor) and Brown University (as Visiting Professor). In 2022 he moved full time to Brown: first as Associate Professor (2022-2025) and currently as Full Professor.

His research is currently funded by the National Science Foundation through several grants (DMS-Analysis, NSF-FRG and NSF-AIMing, together ~2.25M USD), bridging mathematics and machine learning. He has also received an ERC Starting Grant (~1.5M EUR) and an NSF CAREER grant (~600k USD) and is co-PI of the research center grant (endowed with 16.5M USD) of the Institute for Computational and Experimental Research in Mathematics (ICERM).

---

Location

Leipzig University
Paulinum, Augustusplatz 10, 04109 Leipzig
5th floor, Felix Klein lecture hall (P 501)