Scholar
Mislav Balunović
Google Scholar ID: fxkgmGwAAAAJ
ETH Zurich
Citations & Impact
Citations
2,255
H-index
23
i10-index
25
Publications
20
Co-authors
13
list available
Publications
Resume (English only)
Academic Achievements
- Published 'MathArena: Evaluating LMs on Uncontaminated Math Competitions' on arXiv, 2025; 'Proof or Bluff? Evaluating LLMs on 2025 USA Math Olympiad' also on arXiv, 2025; 'MathConstruct: Challenging LLM Reasoning with Constructive Proofs' accepted by ICML 2025; AgentDojo project won first prize in SafeBench competition (50,000 USD).
Research Experience
- Currently a Senior Research Scientist at Google DeepMind; previously worked on securing AI agents at a startup; interned at Twitter, Facebook, and SigOpt.
Education
- PhD from ETH Zurich, advisor unknown.
Background
- Senior Research Scientist at Google DeepMind. His research interests include creating benchmarks like MathArena to evaluate LLMs on the latest math competitions. He received his PhD from ETH Zurich, where his work was featured in media such as Ars Technica, Forbes, and Wired. During high school, he won a gold medal at IMO and silver medals at IOI.
Miscellany
- Released matharena.ai, a website for evaluating LLMs on latest math competitions.
