Niramai — Thermalytix
AI + thermal imaging for early, non-invasive breast-cancer screening
From August 2025 to August 2026, I was a Research Engineer at Niramai Health Analytix (Bengaluru), working under Dr. Geetha Manjunath on Thermalytix, an AI-powered breast-cancer screening solution that combines thermal imaging with machine learning. The approach is radiation-free, non-contact, and painless, and works across women of all ages and breast densities, making early screening more accessible.
I worked on two projects:
- Vision foundation model (R&D). I led continued self-supervised pretraining of a medical vision foundation model (DINOv2) on clinical thermal images, adapting the DINO recipe to the thermal modality with modality-specific augmentations and weighted sampling for class imbalance.
- Edge engineering and hospital deployment. In collaboration with Intel, I led the migration of the clinical imaging software from cloud infrastructure to standalone edge hardware, and presented the work at the Intel Client Ecosystem Symposium 2026 in Taipei. I also carried out the on-site deployment of the screening platform at a public super-speciality hospital near Delhi, working with clinicians, technicians, and hospital IT teams.
🔗 Learn more: niramai.com