Two programmes, one method

Netra: cataract screening and eye-care guidance, on any phone

Cataract is the single largest cause of blindness in India, and most of it is treatable. The gap is not surgery — it is reaching a specialist early enough. Netra closes that gap with a phone camera and a conversation.

A person photographs their eye, describes the problem in their own words, and gets a clear answer: what the signs suggest, how urgent it is, and where the nearest government or subsidised eye hospital is. Free to use, and built to work on low-end phones and weak networks.

• Deep-learning classification of cataract from smartphone eye images
• Conversational triage in plain language, for people with no clinical vocabulary
• Red-flag detection that escalates urgent cases instead of reassuring them
• Referral to the nearest public or low-cost eye hospital
• Designed for Indian-language support and offline-tolerant use

Netra screens and directs. It does not diagnose, and every result points the person toward a qualified ophthalmologist.

End-to-end AI drug discovery: from target identification to molecule

Most platforms own one step. We run the whole chain in-house — picking the target out of sequencing data, then generating and ranking molecules against it — which means the target and the chemistry are optimised for each other rather than handed over a wall.

The method is published, not promised: our receptor-aware peptide diffusion model appeared in J. Chem. Inf. Model., and our single-cell target discovery work in Scientific Reports.

• Targets surfaced from large-scale single-cell and bulk transcriptomics
• Generative diffusion models for receptor-aware peptide and small-molecule design
• Affinity and stability prediction to rank candidates before any wet-lab spend
• Dynamical modelling of pathways to check a target is worth hitting at all