Loading project repository...

PROJECTS

$ ls -la /projects/*

Call Score Analytics

PRODUCTION
Enterprise-grade AI system that analyzes sales conversations in real-time, generating comprehensive transcripts and performance metrics using advanced NLP and GPT-3.5 Turbo. The system processes audio streams, applies conversation intelligence algorithms, and provides actionable insights for sales teams.

Key Features:

  • Real-time audio transcription with 95% accuracy
  • Sentiment analysis and emotion detection
  • Automated call scoring based on custom criteria
  • Integration with CRM platforms
Python OpenAI GPT-3.5 LangChain Apache Kafka WebSocket Docker

Real-Time AI Assistant

PRODUCTION
Revolutionary RAG-based system delivering contextual suggestions during live sales calls with sub-second latency. Leverages vector databases for semantic search and provides real-time recommendations through optimized WebSocket architecture.

Key Features:

  • Sub-second response time using vector search
  • Context-aware suggestions based on conversation flow
  • Integration with knowledge bases and documentation
  • Multi-language support
Python WebSocket RAG ChromaDB Redis FastAPI

Bhagavad Gita AI Chatbot

OPEN SOURCE
Interactive spiritual companion enabling natural conversations with Bhagavad Gita characters using advanced embeddings, ChromaDB, and Google Gemini AI. The system maintains conversational context and provides philosophically accurate responses.

Key Features:

  • Character-based conversations (Krishna, Arjuna)
  • Context-aware philosophical responses
  • Multi-turn dialogue management
  • Citation of relevant verses
Python ChromaDB Google Gemini Embeddings Streamlit

Smart Bus Tracking System

PRODUCTION
Intelligent transportation platform processing 1M+ GPS records daily with ML-powered ETA predictions and microservices architecture. The system provides real-time tracking, route optimization, and accurate arrival time predictions for public transportation.

Key Features:

  • Real-time GPS tracking for 10,000+ buses
  • ML-based ETA prediction with 90% accuracy
  • Route optimization algorithms
  • Passenger mobile app with live updates
Java Microservices RabbitMQ Firebase ML Kubernetes

Article Recommendation Engine

ARCHIVED
ML-powered recommendation system processing 500,000+ articles daily with advanced web scraping, NLP analysis, and personalized content delivery. The engine learns user preferences and provides highly relevant content recommendations.

Key Features:

  • Distributed web scraping at scale
  • Content classification using NLP
  • Collaborative filtering recommendations
  • Real-time personalization
Node.js Python BeautifulSoup ML Redis MongoDB