π About Me I'm passionate about building intelligent, production-grade AI systems that solve real-world challenges. My focus is on the complete development lifecycle, from designing complex agentic workflows with LangChain and LangGraph to deploying scalable voice AI systems using modern tools like LiveKit. I enjoy working in fast-paced environments and turning innovative ideas into tangible, high-impact products.
π§ Core Competencies Generative AI & RAG: Expertise in designing and implementing advanced Retrieval-Augmented Generation (RAG) pipelines, prompt engineering, and building autonomous AI agents.
Conversational & Voice AI: Skilled in creating real-time, intelligent voice agents, integrating STT/TTS services, and performing post-call analysis using tools like Azure Cognitive Services.
Full-Stack AI Development: Proficient in connecting powerful AI backends with intuitive front-end interfaces (ReactJS) to deliver seamless, end-to-end user experiences.
π οΈ My Technology Stack π€ AI & Machine Learning
βοΈ Backend, Frontend & DevOps
β¨ Featured Projects
A real-time, conversational AI voice assistant built with a sophisticated RAG pipeline. It uses LiveKit for real-time communication, LlamaIndex for intelligent document retrieval, and Zilliz Cloud for scalable vector storage. This agent can hold context-aware conversations based on a custom knowledge base.
Tech Used: Python, LiveKit, LlamaIndex, OpenAI GPT-4o, Deepgram (STT), ElevenLabs (TTS), Zilliz Cloud
View Repository β
A full-stack AI system for managing defense fleets, featuring GPS tracking and predictive maintenance alerts. The system uses Gemini AI for predictive analytics and is integrated with Arduino for hardware communication.
Tech Used: ReactJS, Supabase, Gemini AI, Arduino, Map APIs
(Link to repository if public)
An autonomous AI agent system designed to discover, process, and provide structured information about sports tournaments. This system uses LLMs to dynamically generate search queries, scrape web pages for data, and organize the findings into a clean, queryable format served via a FastAPI backend.
Tech Used: Python, FastAPI, LLMs, Agentic Workflows, Web Scraping
View Repository β
π GitHub Statistics