The Beginning: A Personal Mission
Ravi Kumar was six months into his role as a Machine Learning Engineer at Google Bangalore when he received a phone call that changed everything. His grandmother, living in a village outside Nagpur, had been misdiagnosed with a common cold — when she actually had early-stage tuberculosis.
"The nearest qualified radiologist was 80 kilometers away," Ravi recalls. "By the time we got a proper diagnosis, she had been suffering for three months. That's when it hit me — AI can read X-rays faster and more accurately than most human readers. Why isn't this technology in every primary health center?"
Six weeks later, Ravi resigned from Google. His friends thought he was crazy. His parents were terrified. But Ravi had a clarity of purpose that, as he puts it, "made the decision feel inevitable."
Building MedAI: The First 100 Days
Ravi started MedAI from his apartment in Indiranagar, Bangalore, with ₹12 lakhs in personal savings and a dataset of 50,000 chest X-rays from a partnership with CMC Vellore. His first hire was Deepa Rao, an IIT Madras PhD in computer vision who shared his conviction that AI could democratize healthcare.
"We didn't want to build another AI tool for affluent urban hospitals. We wanted to build for the PHC in Yavatmal, the district hospital in Chhindwara. That's where the impact is." — Ravi Kumar, CEO, MedAI
The technical challenge was formidable. Most AI diagnostic models are trained on Western patient data — they perform poorly on Indian demographics. MedAI's breakthrough was building a model trained exclusively on Indian patient data, accounting for differences in body composition, disease prevalence, and X-ray equipment quality.
The Pivot That Changed Everything
By month eight, MedAI had a working product — a mobile app that could analyze a chest X-ray in under 30 seconds and flag potential TB cases with 94.2% accuracy. But they hit a wall: rural health workers didn't use smartphones. The X-ray machines were old analog devices. The internet connectivity was spotty at best.
"We had built a Silicon Valley product for rural India," Ravi laughs. "Classic mistake." The team pivoted to a hardware solution — a ₹15,000 portable AI device that could be attached to any X-ray machine, process images offline, and send results via SMS.
Impact at Scale
Today, MedAI's devices are deployed in 847 primary health centers across Maharashtra, Madhya Pradesh, and Tamil Nadu. They've screened over 2.3 million patients, flagging 142,000 for further examination — of which 38,000 were confirmed TB cases that would have gone undiagnosed.
The numbers are staggering: MedAI estimates that their early detection has prevented an estimated 12,000 TB transmission chains, potentially saving tens of thousands of lives.
What's Next
With their recent ₹83Cr seed round led by Sequoia Scout, MedAI is expanding beyond TB. They're training models for malaria, pneumonia, and cardiac conditions — all optimized for Indian patient demographics and low-resource clinical settings.
Ravi's vision is audacious: "By 2028, every primary health center in India should have AI-assisted diagnostics. Not as a luxury — as a standard. The technology is ready. The infrastructure is ready. We just need the will."