Drug Safety Classification System (RAG Pipeline)
Identified critical accuracy gaps in Tata 1mg's drug safety classification system.
Led the design of a retrieval-augmented generation (RAG) pipeline to automate validation and classification of drug safety advisories.
The system validates clinical content and regulatory classifications across safety categories like pregnancy, breastfeeding, driving, and organ impairment.
The Challenge
- Manual classification against official drug labels was producing 48% safety advice accuracy, meaning over half of the drugs had incorrect or incomplete safety details visible.
- In a clinical context, an incorrect pregnancy or breastfeeding classification is a patient safety risk, not merely a content quality issue.
- Diagnostics over-claims and incorrect assertions were creating regulatory and brand risk.
The Approach & Solution
- Scoped and co-designed a RAG pipeline that retrieves the relevant section from the official drug label (SmPC), compares it against Tata 1mg's internal classification, and maps the correct advisory.
- Defined the evaluation rubric and accuracy benchmarking framework for the pipeline's output validation.
- Collaborated with engineering and medical affairs to establish the confidence threshold for automated vs human-reviewed classifications.
Key Results & Commercial Impact
Technology Stack & Tools
Have a similar project in mind?
We specialize in custom AI systems, clinical data licensing pipelines, and scalable web apps. Let's build something exceptional together.
Explore More Projects
Google Health Partnership
Packaged clinical content as LLM training data, including drug monographs, DDI data, disease content, and regulatory safety data, in a B2B monetization deal.
Samsung Health API Integration
Real-time API integration spanning 600K+ SKUs, 50+ attributes, and a 200K-interaction DDI tool, enabling in-app discovery and routing cart/checkout to Tata 1mg.
Elsevier Clinical Monograph API Partnership
Tata 1mg's first clinical content licensing deal. Delivered 1,500 structured drug monographs over 9 months and built the CMS & API infrastructure.