Prabhas Reddy

At National Small Industries Corporation (NSIC), I built ResolveAI, a multi-agent customer support system using Python, LangGraph, LangChain, and FastAPI. It classified complaints with 87% accuracy and integrated agents for policy retrieval and automated complaint resolution.

I also built ResolveAI as a project, developing a RAG pipeline with ChromaDB to retrieve relevant policies and generate context-aware responses. LangGraph coordinated agents for complaint handling, order verification, and resolution generation, with human escalation for complex cases.

For my Mental Health Support Chatbot, I built text and voice conversations using RAG over DAIC-WOZ clinical transcripts. It included PHQ-8 assessments, voice-tone adaptation, conversation history, and crisis-resource support; in a 15-participant evaluation, it earned 4.3/5 for perceived empathy and 4.1/5 for contextual relevance.

I also built DataPilot, a natural-language-to-SQL application that validates and executes queries, handles errors, and automatically corrects queries. Its SQL query execution accuracy was 89.2% in testing, and its React interface lets users explore databases, visualize results, and generate AI-powered insights.