IndicAgri

Multi-modal RAG chatbot supporting 20+ Indian languages for agricultural advisory. 🏆 Top 14 Nationals – Capital One Launchpad Hackathon

Top 14 NationalsCapital One Launchpad Hackathon (Aug 2025)

  • Developed multi-modal Retrieval-Augmented Generation (RAG) chatbot supporting 20+ Indian languages for agricultural advisory services.
  • Engineered autonomous data pipeline constructing 15,000+ document knowledge base from government agricultural resources; implemented chunking strategies and embedding optimization for domain-specific retrieval.
  • Achieved 45% reduction in inference latency through model quantization (INT8/INT4), KV-cache optimization, and intelligent response caching mechanisms.

View on GitHub