Researcher at ETH Zürich & Max Planck Institute for Intelligent Systems
Co-author of the research paper on test-time curricula for reinforcement learning
How media typically covers Ido Hakimi
Ido Hakimi as author
Test-time curriculum reinforcement learning improves model performance on target tasks without human dataset curation, achieving 1.8x pass@1 improvement on AIME25 and 2.1x on CodeElo for Qwen3-8B.
“Co-author of the research paper on test-time curricula for reinforcement learning”