Clinical intuition doesn’t scale. We can build the systems that do.
I spent years inside the NHS, running ward rounds, making decisions under uncertainty at 3 am, watching brilliant clinicians bottlenecked by systems that were never designed for them. The failure mode of modern healthcare isn’t a lack of talent. It’s that the hard part of healthcare AI, isn’t the AI. It’s knowing which clinical problems are actually worth solving, and which ones just look good in a pitch deck. I didn’t leave medicine because I lost faith in it. I left because the highest-leverage thing I could do for patients was to stop treating them one at a time.
So I crossed over. I was Employee #1 at a Khosla Ventures-backed AI startup that hit an $18M valuation. I advised a $200M+ venture fund on healthtech due diligence, telling investors which founders actually understood clinical reality and which were cosplaying. I trained at Imperial, the top 1% of my management cohort, and obtained my MBBS, because I never saw clinical and commercial problems as separate. Now, as COO of Rhazes AI, I’m building an AI co-pilot that thinks with doctors, not instead of them.
Most founders in this space have seen a hospital from a pitch deck. I’ve seen one from the inside of a 12-hour night shift. I know what it feels like to document the same note in different systems while a patient waits. And I know what it feels like to watch an engineering team build something elegant that no doctor will ever use, because no doctor ever asked for it. That gap is where most healthtech products go to die, and it’s exactly where I’ve chosen to build.
Core Convictions
The Domain Expert Is the Moat
In healthtech, your unfair advantage isn't your model architecture; it's the founder who has personally misdiagnosed a patient at 2 am and understands why the system failed. LLMs are a commodity. Clinical judgment is not. The startups that will win are those in which the builder and the end-user share a brain.
Build for the Workflow, Not the Demo
The graveyard of healthtech is full of products that looked incredible on stage and collected dust in hospitals. Doctors don't adopt tools because they're clever. They adopt tools that save them time between patients. If your product adds even thirty seconds of friction to a clinical workflow, it's game over. I've stood on both sides of that door, and I build accordingly.
Healthcare Needs Founders, Not Tourists
Too much capital flows into healthtech from people who see a $4 trillion market and think the hard part is the TAM slide. It's not. The hard part is understanding why a consultant won't trust your algorithm, why an NHS trust takes eighteen months to procure anything, and why regulatory isn't a checkbox; it's the product. I'm not visiting this industry. I come from it.
Building in Public
Documenting the real journey — clinical AI, venture capital, founder life — one short-form video at a time.