ShinkaEvolve, Evolved: Faster, Cheaper, in Your Coding Agent

Robert Lange is a Staff Research Scientist and founding member at Sakana AI, where he works on nature-inspired approaches to AI, automated discovery, evolutionary computation, and self-improving systems. His research sits at the intersection of large language models, evolutionary search, meta-learning, and scientific automation. Across projects such as The AI Scientist, ShinkaEvolve, EvoLLM, and learned evolutionary optimization, his work explores how AI systems can generate, test, improve, and iterate on ideas with less human hand-holding. Before Sakana AI, Robert built a research background spanning economics, data science, computing, and evolutionary meta-learning, with academic work connected to TU Berlin and prior internships at Google DeepMind. His broader focus is on making AI systems more open-ended: systems that can search over methods, improve their own tools, and help discover better algorithms rather than simply execute fixed instructions.