ZaynichOne
A production React Native app for Wockhardt, a pharmaceutical MNC, featuring an agentic RAG chatbot that answers drug-related questions from an approved knowledge base.
The Problem
Staff and users needed fast, accurate, grounded answers to drug-related questions without digging through documents, in a setting where accuracy and privacy are critical.
What I Built
As GenAI engineer and app developer, I built both the app and the AI behind it.
- The full ingestion pipeline that processes source pharma documents into a knowledge base
- The agentic RAG system that grounds answers in approved sources
- Privacy and data-handling boundaries suited to a regulated pharma setting
- The React Native mobile app across iOS and Android
The Hard Part
Keeping answers grounded and accurate on drug data, where a wrong answer is unacceptable, and building retrieval that stays reliable rather than hallucinating.
Outcome
Live and released on both app stores for a real pharmaceutical company.
Tech Stack
- React Native
- Python
- LLMs
- RAG
- Agentic Workflows
