India Builds AI That Keeps Patient Data Private and Local

Healthcare systems across India hold millions of patient records that could save lives if analyzed together, but privacy concerns have kept that data locked away in separate hospital systems.

Now, a Pune-based AI company has found a way to break through that barrier. Partex.AI has developed India's first sovereign healthcare foundation model that can learn from patient data across multiple hospitals without ever transferring a single medical record.

Gaurav Tripathi, Co-founder and Group CTO at Partex.AI, shared the breakthrough at DevSparks 2026 in Pune. He explained how the system flips traditional AI training on its head: instead of sending sensitive patient data to a central computer, the AI model travels to each hospital and learns locally.

"We are not moving the data to the model. We are bringing the model to the data," Tripathi said. Only encrypted updates get shared between systems, meaning patient records never leave their original hospital.

The approach, called federated learning, solves a problem that has frustrated healthcare researchers for years. Doctors know that analyzing patterns across thousands of patients could help detect diseases earlier and improve treatments, but strict privacy laws prevent hospitals from pooling their data.

The Bright Side

Partex's system does more than just protect privacy. It's designed specifically for India's unique healthcare landscape, supporting multiple languages and accounting for disease patterns that differ significantly from Western countries.

The model uses knowledge graphs to map complex relationships between genes, diseases, and treatments in ways that reflect Indian medical realities. This means the AI understands connections that global models trained primarily on European or American data might miss.

The breakthrough arrives as India implements its Digital Personal Data Protection law, which places strict limits on how patient information can be shared. Tripathi emphasized that trust matters as much as technology when handling medical records that involve patients, hospitals, research institutions, and pharmaceutical companies.

The system could help clinicians spot disease trends across populations, understand which treatments work best for different patient groups, and develop therapies tailored to Indian healthcare needs. All while keeping every patient record exactly where it belongs: safe, private, and under local control.