Medical AI
Foundation models and decision support that reason over sparse, real-world health data, from wearable signals to clinical risk prediction.
We’re building AI to turn every doctor into a specialist. The right treatment, for the right patient, at the right time.
AI is accelerating biotech. As treatment possibilities expand, the bottleneck is increasingly the limited number of specialists responsible for determining which of these new treatments will be right for each patient, and when.
Mirae is solving this information bottleneck by connecting the complete context of each patient to the knowledge required to make the right treatment decision, making specialist-level care available wherever the patient is.
Working alongside the Computational Health Informatics Lab at the University of Oxford.
Foundation models and decision support that reason over sparse, real-world health data, from wearable signals to clinical risk prediction.
Methods for learning from health records without centralizing them: federation, condensation and abstraction over heterogeneous data.
2026 · Nature Machine Intelligence
Pretrained on data from 1.7 million individuals, the model assesses cardiac health from everyday device signals, the kind of continuous, between-visit data Mirae is built on.
Read paper2025 · medRxiv
An agentic system that selects and runs validated clinical risk tools.
Read paper2023 · arXiv
Explores dataset condensation to create compact representations of clinical records for more efficient, privacy-conscious AI research.
Read paperCo-founder & CEO
Grew PatientsLikeMe from $0 to $16M ARR in 3 years as Founding GTM hire. Founded and grew Motto Health to $1M ARR. Built health-tech ventures at iCarbonX and Roivant.
Co-founder & CTO
Co-founded mPharma ($100M raised), scaled it to 1,000 employees and serving millions of patients every year across Africa. Led AI and engineering at GitStart (YC S19), building SWE agents for clients like Google.
Academic Co-Founder
Royal Academy of Engineering Chair of Clinical Machine Learning at Oxford University. Top 1% medical AI researcher worldwide (G. H-index of 86, 400+ peer-reviewed papers). Supervised 25 Rhodes Scholars and talent that joined OpenAI and Anthropic.
Founding Engineer
National Maths Olympiad winner, Microsoft TypeScript engineer. Founded GitStart (YC S19), raising $10M and serving clients including Google.
Founding Engineer
Former mPharma CTO, where he led a team of more than 50 engineers.
Founding Engineer
Wrote the most downloaded React Native library for Apple HealthKit worldwide, connecting mobile apps with health data.
Founding Designer
Spent 11 years designing products at Grubhub as it grew to $2B in revenue and 25M monthly active users.
Founding Ops and Finance
Former IBM finance professional and investor for eight years. Brought two products to market as a founder and CTO.
Postdoctoral Researcher
Researches vision-language models and trustworthy machine learning for healthcare. Published work in Nature Communications, ICLR, NeurIPS, and WWW.
Senior Research Associate
Previously researched at Meta and Microsoft Research Asia. Co-leads Oxford's Digital Health Group in Suzhou.