Hi, I'm Ravindra Yadav
An AI & Full-Stack Developer building production-ready RAG systems, multi-LLM platforms, and real-time apps with React, Next.js, and Node.js.
ChatMate
A ChatGPT-style multi-LLM chat platform with real-time streaming, conversation branching, and tool calling across OpenAI and Gemini.
ChaiBookLM
A full-stack RAG research assistant inspired by Google NotebookLM — upload sources, build a knowledge base, and query it conversationally.
MentorOS
An AI mentoring platform with dual mentor personas, YouTube tool integration, and multi-provider LLM routing with rate-limited access.
AI Arena
A side-by-side LLM comparison tool with concurrent SSE streaming across providers and a structured judge-model scoring pipeline.
ConvoX — Real-Time Chat Application
A real-time chat application with a modern UI, JWT auth, friend requests, and Redis-backed scalability.
Where RAG Fails: Understanding the Limitations
A practical look at why Retrieval-Augmented Generation breaks down in real applications — chunking pitfalls, retrieval blind spots, and what actually helps AI answer questions from your own documents.
Context Window — Humans vs. LLMs
How memory actually works for large language models compared to how humans remember, and why understanding context windows changes how you should prompt and design AI applications.
How ChatGPT Understands Your Questions
Tracing the path from raw text to intelligent response — tokenization, embeddings, and attention explained simply, for anyone curious what's really happening under the hood.