Harri Valpola

Curious AI builds industrial automation AI systems based on predictive world model approaches to machine learning, applying research in predictive coding and unsupervised representation learning to industrial control and automation applications. The Helsinki company is founded by Harri Valpola, a prominent Finnish machine learning researcher who pioneered variational autoencoder approaches and has published extensively on brain-inspired AI architectures.\n\nThe company raised approximately $10 million in Series A funding from Finnish and European investors. Curious AI targets industrial automation use cases where sensor data from manufacturing environments can train predictive models that anticipate equipment behaviour, enabling more robust and adaptive control systems than traditional programmed automation. The company has worked with Finnish industrial companies on proof-of-concept deployments of predictive world model control systems.\n\nCurious AI occupies a niche in the industrial AI market characterised by fundamental research ambition combined with practical automation application targets. The company founder academic profile distinguishes it from most industrial AI startups, as Valpola contributions to unsupervised learning and predictive coding represent genuine theoretical contributions to the field. Finland strong manufacturing and engineering industry, including major companies in paper, energy, and electronics, provides a testbed for industrial AI applications, though the company competes for industrial automation contracts with much larger vendors including Siemens, ABB, and Rockwell Automation that have significantly greater distribution and industry relationships.