# About Me

I'm an ophthalmic surgeon on secondment from full-time surgical training in the North London Deanery.

Currently, I'm working full-time on conversational AI safety, serving as Science Director at Ufonia, where I coordinate the clinical, regulatory, and safety aspects of Dora — our autonomous clinical conversational assistant.

I'm also a final-year Computer Science PhD student at the University of York, focused on assuring conversational AI agents for healthcare. (More on my research below.)

I lead a team of product managers and machine learning engineers to build, evaluate, and deploy AI agents that can safely support real clinical conversations. I also lead our portfolio of >£2.5m of research and development grants awarded by Innovate UK and NIHR as Chief Investigator, and help support our regulatory strategy for the UK and US.

One of the things I'm most proud of is setting up and keeping our company's AI research journal club alive for over 2 years - bringing together clinicians, engineers, and researchers to dive into the latest in AI safety, LLMs, and clinical AI.

I'm also a:

• 💻 Weekend programmer turned hobbyist coder – I taught myself to code via Harvard's CS50 and now build side projects in my spare time. Link to projects
• 🎓 Alumnus of the Value-Based Healthcare Intensive Seminar at Harvard Business School
• 🚀 NHS Clinical Entrepreneur

I also find teaching and mentorship extremely satisfying:

• 👩‍💻 Module lead on UCL's Bioscience Entrepreneurship MSc, where I teach how to bring digital health products to market
• 👁️ Founder of stinterview.com, an interview prep network for aspiring ophthalmologists
• 🛠️ Creator of Clinician2Creator, my course that helps clinicians learn to build and ship software products
• 🧪 Mentor on the Bitelabs Fellowship for clinicians looking to move into tech.

My research at Ufonia and the University of York explores how to safely implement conversational AI in real clinical workflows, drawing on safety principles from aviation, nuclear, and other high-integrity domains.

I began by leading clinical evaluations of AI agents in ophthalmology as medical director at Ufonia (1, 2, 3).

With the rise of LLMs, my focus has shifted to understanding their risks (4) and capabilities (5), and how they can be safely embedded in patient-facing systems for replacing parts of clinical workflows.

I work closely with our engineering team to build scalable evaluations that:

• Provide verification and validation evidence to support regulatory submissions
• Guide iterative product development throughout the lifecycle

A key theme across my work is how to evaluate and assure non-deterministic systems. See our recent framework: [6]. I believe that robust, validated evaluations situated within structured safety arguments are the key to safely deploying AI in healthcare.

Our work has been presented at leading clinical conferences such as ARVO, ESCRS, AAO, and technical conferences such as NEURIPS and ACL.

• 🍳 Cooking – Keen home cook, especially drawn to Mediterranean and Korean cuisine
• 🎾 Tennis – Picked it up recently and completely addicted
• 📚 Reading – Love historical fiction, science fiction (Culture series, Asimov), and biographies