Multi-agent system routing requests across weather, Gmail, translation, web search, and math. Intent-routing, planning, memory, and self-evaluation with ChatGroq LLM.
RAG chatbot running semantic search, knowledge graph, and BM25 simultaneously — picks the highest-accuracy answer. Ingests text, PDF, and structured data dynamically.
This project focuses on identifying barren or underutilized land suitable for vegetation restoration in both plains and hilly regions using satellite imagery and Google Earth Engine. The system analyzes geographical and environmental factors such as slope, steepness to determine areas with high restoration potential
- Sole AI developer — owned the complete Generative AI solution end-to-end
- Built real-time voice AI interviewer using Web Speech API and Google Gemini
- Designed prompt flow, conversational logic, and contextual response handling
- Collaborated with frontend developers on HTML/CSS/JS implementation