Research, releases, and updates from the rLLM Project — building the infrastructure to train, evaluate, and evolve intelligent agents.