The work
I trained as a doctor in India, graduating from Kasturba Medical College in Manipal in 2019. I joined the medical operations and product team at DailyRounds/Marrow in Bengaluru, where I spent two and a half years helping scale India's leading medical education platform from an early-stage startup to over half a million users. I led a cross-functional team of twenty medical editors, designed the flagship "Pearls" feature, and did the unglamorous day-to-day work of translating educational needs into product requirements.
Between 2022 and 2023 I ran a small family practice in Kerala. It was paper-based when I arrived. I partnered with a fintech-backed EHR startup to install and configure a digital record system; within weeks the clinic was running on software that our staff actively preferred to the old ledgers. That experience — being the only person in a resource-limited clinic who understood both the medicine and the technology — shaped how I think about health-tech deployment.
I moved to London in late 2023 to practise emergency medicine at Whipps Cross Hospital as part of Barts Health NHS Trust. I'm currently an ST3-equivalent trust-grade doctor in a busy adult A&E, covering the full range of undifferentiated presentations across resus, majors, and ambulatory care. I co-lead a departmental quality improvement project on mental health pathway compliance and I teach medical students and foundation doctors on shift.
The building
Alongside clinical work, I build things. Most of them clinical AI.
I've been experimenting with machine learning for healthcare since 2018, when I trained a convolutional neural network in PyTorch to prioritise chest X-rays as a final-year student. In 2020 I placed second at the MIT COVID-19 Hackathon with a team building AI triage tools. In 2022 I launched NeurallyMed — an AI-powered PLAB exam preparation platform — which was acquired in 2023.
I'm a founding engineer at two sister ventures, Peerr.io and Aiigent.io, working on secure, AI-powered tooling for clinical networking and workflow optimisations. On the side I build smaller things I need — an AI guidelines retrieval platform for the ED (Aide), a fitness and nutrition assistant for myself (Leanifi), a study tool I built while sitting the MSRA (msra.uk), a project microsite for the QIP I'm running at Whipps (qip.fazeen.dev). Some of these turn into products. Some stay personal tools. All of them start with a problem I saw on a shift.
What I care about
Most clinical AI projects fail not because the model is weak but because the implementation is weak. Poor workflow integration, weak governance, unclear accountability, thin evaluation. I want to do the opposite — ship things, evaluate them properly, write down what worked and what didn't so other people don't have to learn the same lessons from scratch.
That's the thread.
Available for collaborations in clinical AI deployment, digital health product consulting, and speaking on clinician-led technology development. Contact me.