About me
I am co-founder and CTO of Klaris, where we build AI agents that help MedTech companies get through regulatory review faster and with fewer surprises. Regulatory frameworks for medical devices are, at their best, structured, logical, and clearly aimed at patient safety. In practice, manufacturers often get stuck proving consistency across hundreds of pages of technical documentation and checking whether every applicable requirement was addressed — work that crowds out the judgment calls that actually matter for safety. Klaris exists to take on that administrative burden so regulatory teams can spend their attention where it counts. As CTO I own the product and technical direction end to end, from the AI systems we ship to the team, processes and security posture behind them.
What draws me to this problem personally is how neatly it sits at the intersection of two long-standing interests of mine: language, and multi-agent systems. Much of my earlier research was about agents that reason, interact and negotiate through language; autonomous driving later sharpened that into planning and prediction under uncertainty in a safety-critical world. MedTech regulation asks for a similar combination — careful reading of complex text, structured reasoning over requirements, and systems that must be trustworthy when the stakes are high.
Klaris grew out of my time as an Entrepreneur in Residence at Antler (autumn cohort 2024, London), where I partnered with Francesco Corazza, completed the incubator programme, and secured funding to start the company.
Previously, I was Lead Research Scientist at Five AI / Bosch (to July 2024), leading Motion Planning and Prediction applied research for autonomous driving. My work spanned multi-agent interaction, robust planning, integrating prediction with planning, and later HD map reconstruction from sensor data — hybrid combinations of deep learning and classical methods in a safety-critical setting.
Before industry, I completed a PhD in the ILCC at the School of Informatics, University of Edinburgh, supervised by Alex Lascarides and Subramanian Ramamoorthy.
Interests
- Generative AI: Agentic systems, Retrieval-Augmented Generation, LLM applications
- Machine Learning: (Multi-agent) Reinforcement Learning, Imitation Learning, Deep Learning, Bayesian Methods
- Prediction, Planning and Decision Making; Computer Vision
- Multi-agent Systems and Game Theory
- Applications in MedTech regulatory compliance and Autonomous Driving
Education
- PhD Informatics. Low-resource learning in complex games, School of Informatics, University of Edinburgh, 2018
- BSc (Hons) Robotics with Artificial Intelligence, University of Bradford, 2013 (1st class);
Awards
- Received Best student Overall Performance Award, University of Bradford, 2013
- Received Best Student Paper Award for “Exploiting action categories in learning complex games”, 2017