David Ha

David Ha is an AI researcher and entrepreneur whose work spans generative models, reinforcement learning, and approaches inspired by biological systems. He studied engineering at the University of Toronto and earned a doctorate at the University of Tokyo. Before becoming known for AI research, he worked in finance, including a senior role at Goldman Sachs.

Ha later worked at Google Brain, leading research in Japan. With Jürgen Schmidhuber, he published World Models in 2018, investigating how learned representations of an environment could support an agent's behavior. The work combined a model of observations with a compact controller and became a recognizable example of research on learned internal models.

After a period leading research at Stability AI, Ha co-founded Sakana AI in Tokyo in 2023 with Llion Jones and Ren Ito. He serves as its chief executive officer. The laboratory explores evolutionary approaches, model combinations, and research automation, drawing inspiration from the ability of natural systems to adapt and cooperate.

His career connects research experimentation with the organizational responsibilities of building a laboratory. Rather than treating progress solely as the training of ever-larger individual networks, Sakana's research investigates ways to discover useful combinations and procedures. Ha's contribution includes published technical work as well as the creation of a setting for exploring these alternative paths to AI development.