Recursive AI has developed BUILD, a system that automates scientific discovery in AI. Its proprietary approach explores ideas beyond conventional algorithm design, with the goal of making algorithm invention scalable across artificial intelligence.
Get in touchAlgorithm development has traditionally depended on the insight of a small number of specialists. BUILD turns it into a systematic, automated process of invention, which vastly speeds up algorithmic innovation. BUILD works at the level of the algorithms underlying AI, and its results apply across many models and applications.
BUILD is a system that automates scientific discovery in AI. It invents new algorithms through a proprietary discovery process that finds ideas outside the usual paths of algorithm design and turns them into measurable improvements in accuracy, reliability, or efficiency.
BUILD can be applied across many areas of AI, including language, vision, continual learning, forecasting, retrieval, and decision systems. The long-term goal is to make the invention of better AI algorithms scalable.
Recursive AI LLC is developing BUILD to automate scientific discovery in AI. Its proprietary discovery process explores unconventional paths to inventing new algorithms across a wide range of AI fields. Early research has produced nine algorithms with experimentally verified improvements across eight areas of AI, nine research manuscripts, and three international conference acceptances. The goal is to turn this discovery capability into a scalable business.
William Claster is the founder and CEO of Recursive AI. His background combines building commercial software businesses with developing machine-learning algorithms.
He founded KK AbleCommerce Japan, which developed e-commerce technology for the Japanese market, became an official supplier to Hitachi Information Systems, and served customers including Marubeni and Shop Japan. He later founded ChatMiner, a social-media intelligence platform, and founded InsiderMiner, where he served as CTO and chief scientist, developing machine-learning systems for forecasting financial markets and delivering predictions through a mobile application.
He also developed fraud-detection models at Paidy, the Japanese financial technology company later acquired by PayPal for approximately $2.7 billion.
At Recursive AI, he leads the development of BUILD, defining its research direction, experiments, and technical strategy. He is also an Associate Teaching Professor at Northeastern University and Professor Emeritus at Ritsumeikan Asia Pacific University.
We're exploring partnerships and early applications of our technology.
wclaster@recursiveai.online