Year

2026

Tech & Technique

React, Node.js, Pinecone, MongoDB, Transformers.js, LLM, RAG

Description

Built YogaFlow, an AI-powered yoga wellness chatbot leveraging RAG (Retrieval-Augmented Generation) architecture for intelligent, context-aware responses with seamless source attribution.

Key Features:
  • RAG architecture for context-aware knowledge retrieval
  • Real-time safety detection and smart safety pivots
  • Personalized yoga recommendations and guidance
  • Secure query logging and conversation history
  • Contextual LLM responses with source attribution
  • Vector-based semantic search using Pinecone

Technical Highlights:
  • Full-stack AI platform with React frontend and Node.js backend
  • Pinecone vector database for efficient knowledge retrieval
  • MongoDB for persistent storage and user data
  • Transformers.js embeddings for semantic understanding
  • LLM integration for intelligent response generation

My Role

Full-Stack AI Developer
  • Backend: Built RAG pipeline with Pinecone and LLM integration
  • Frontend: Created interactive chat interface with React
  • Safety: Implemented real-time safety detection and smart pivots
  • Embeddings: Integrated Transformers.js for semantic search
  • Database: Designed MongoDB schema for query logging
YogaFlow

VIDHIT