The people closing the loop on self-improvement

Emeritus Professor at the University of Cambridge and pioneer in reinforcement learning for speech and dialogue systems. Founder of multiple successful AI ventures including VocalIQ, acquired by Apple in 2015, and Entropic, acquired by Microsoft in 1999. Steve also served as Chairman of Phonetic Arts, acquired by Google in 2010, and was an early advisor and mentor to conversational AI startup PolyAI, who raised their Series D in December 2025. Steve has spent decades working at the frontier of adaptive AI, intelligent dialogue systems, and machine learning research, and provides strategic and scientific guidance across Recurvia's long-term research direction.

Former Machine Learning Research Engineer at Apple, where he worked on large-scale neural network systems for Siri. John holds MSc and PhD degrees in Artificial Intelligence and Natural Language Processing from the University of Edinburgh, and an undergraduate degree in Linguistics from the University of Cambridge. John co-founded Recurvia in 2024 with the goal of building technologies which safely enable AI systems to analyse, optimise, and improve themselves.

Machine learning researcher and mathematician specialising in reinforcement learning, optimisation, and large-scale AI systems. Prior to joining Recurvia, Tristan was a Staff Research Engineer at InstaDeep, where he worked on reinforcement learning systems for industrial optimisation and commercial AI deployment at scale. Tristan holds a PhD in Mathematics from King's College London and undergraduate and masters degrees in Mathematics from Cambridge.

Enterprise infrastructure and AI commercial specialist with more than a decade of experience spanning software engineering, technology consulting, and enterprise sales. Prior to joining Recurvia, Alexandra worked at Neptune.ai (acquired by OpenAI in December 2025), where she helped leading organisations deploy and scale machine learning infrastructure, partnering with engineering and MLOps teams across industries including pharmaceuticals, defence, finance, and technology. Alexandra holds a degree in Systems Engineering and Computer Science from The George Washington University. She leads the commercialisation of Recurvia's AI-powered optimisation and observability technologies.

Software engineer and entrepreneur specialising in large-scale infrastructure, machine learning systems, and security engineering. James holds a First Class degree in Mathematics from King's College London. Prior to co-founding Recurvia in 2024, he founded and successfully exited Arixian, an AI-centric physical security company. His work focuses on building robust systems capable of supporting next-generation AI workloads at scale.

Elliot holds a PhD in multi-agent reinforcement learning from the University of Edinburgh, with prior research roles at Honda Research Institute and FiveAI spanning multi-agent learning and autonomous systems. His interests run from reinforcement learning and multi-agent systems to dynamical systems, forecasting, and robotics — the mathematics of how complex systems evolve, and how to predict and steer them. At Recurvia, those threads converge: a training run is a dynamical system, and steering it is a multi-agent control problem. Elliot leads experimentation and evaluation across optimisation and observability workloads, driving the development of self-improving training systems at scale.